AVForums South Africa

Audio/Video Hardware => Valves/Tubes => Topic started by: Mars on March 25, 2019, 11:14:14 AM

Title: Mars 5-10 stereo - update
Post by: Mars on March 25, 2019, 11:14:14 AM
Hi All

We had a request to develop a Mullard 5-10 type stereo amplifier kit in 2018. The result was so good, that we have decided to make it our EL84 kit offering of choice. Since then we have had two orders for kits with red transformer covers. I took some photos before shipping. What do you think of the look?

(https://i.imgur.com/q9mZY8a.jpg)

(https://i.imgur.com/2Os4Idx.jpg)

(https://i.imgur.com/F54f7ca.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on March 26, 2019, 10:39:41 PM
That looks excellent Karel . Quite unique . Where's the "Mars" badge ?  ;D

I was wondering .... Are the 5-10's transformers wound with Ultra-Linear taps ? Does the circuit differ from the classic Dynaco 5-10 ? Is there a choke ?

I also think it may be a good idea to encourage Mars builders to "pimp" their amps . It could lead to all kinds of exciting looking gear .
Title: Re: Mars 5-10 stereo - update
Post by: JohnnyIC on March 28, 2019, 08:45:24 AM
What would a kit like this retail for? I have been meaning to build a diy tube amplifier for some time now and this looks simple enough for a noob like myself.
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on March 28, 2019, 10:56:31 AM
That looks excellent Karel . Quite unique . Where's the "Mars" badge ?  ;D

I was wondering .... Are the 5-10's transformers wound with Ultra-Linear taps ? Does the circuit differ from the classic Dynaco 5-10 ? Is there a choke ?

I also think it may be a good idea to encourage Mars builders to "pimp" their amps . It could lead to all kinds of exciting looking gear .

Here is the Original 5-10 (http://www.r-type.org/articles/art-003e.htm) with the circuit diagram below. It looks like the tone controls have been removed, and the transformers aren't ultralinear. But maybe @Mars can tell us how this version differs? Karel do you use the same output transformers as your EL84 PP, or have you wound different ones according to the  Mullard spec?

(http://www.r-type.org/articles/art003bc.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on March 28, 2019, 11:52:20 AM
One difference seems to be a GZ34 (?) rectifier instead of an EZ81

Understandable because is has a bit more oomph
Title: Re: Mars 5-10 stereo - update
Post by: charles on March 28, 2019, 12:35:17 PM
Some interesting information regarding the Mullard 5-10 amplifier

I have an original 5-10 Mullard manual (very early edition).

Rectifier valves GZ30 (5V) or EZ80 (6.3V)

This manual (86 pages) even provides Minimum Specifications for the Output transformer such as:

Matching

Winding data: Laminations, Core Material, Core Size, Winding Table, Wire, With of Windings,
Insulation, Winding Sense and Electrical Characteristics.

I would really like to measure some of the new 5-20 Output Transformer's specifications against the
original Electrical Characteristics given in this manual


   
Title: Re: Mars 5-10 stereo - update
Post by: Nidri on March 28, 2019, 08:57:00 PM
That hammertone finish is really nice.
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on March 30, 2019, 04:41:27 AM
It looks like the tone controls have been removed,....

No, they are still there; the two pots and associated components just before the volume control, leading into the EF86 g1. (Can't quite read the figures and not sure where my original is). It is a passive tone control configuration, in the days where the signal feed would have been some 500mV, and the pick-up mostly a crystal/ceramic cartridge.

The circuit would benefit from U.L. output transformers; such were not universally available at the time. (25% - 33% taps would be sufficient.) The then higher screen voltage would require some re-biasing, unless the h.t. could be brought down to about 330V. (With a 300Vrms power transformer secondary it might be up to 360V).

Some folks might remember that with this circuit there is the general complaint of being too sensitive, particularly with the tone controls removed.

A Mr Claude Byrith have decreased the gain by using the EF86 triode-connected, which circuit is often referred to.  In my opinion , particularly when building a new amplifier, that choice is second best, though I can see the logic when an initial already exists. But that is another story for another occasion. (I do not recall what Karel's own original looked like; I think it uses a cathodyne phase splitter. There is also merit in that. Perhaps we can discuss all this in PM should he be interested. The original 5-10 can be audibly improved in its behaviour above 2 KHz.)


P.S:  El Sid: Sorry, I may have misunderstood! Did you mean that in the Mars kit the tone controls have been removed, not in the schematic itself that you gave?
Title: Re: Mars 5-10 stereo - update
Post by: kenvanraas on March 30, 2019, 04:45:28 AM
Haai oom Amp!  :dance:
Gaan nou zzz,dis 04h44 Chow ne  :dance:
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on March 30, 2019, 08:39:13 AM
P.S:  El Sid: Sorry, I may have misunderstood! Did you mean that in the Mars kit the tone controls have been removed, not in the schematic itself that you gave?

Yes indeed - Im keen to see how the Mars version differs from the original. The Mars version appears to have no tone control knobs hence my observation
Title: Re: Mars 5-10 stereo - update
Post by: Family_Dog on March 30, 2019, 09:03:59 AM
The Mullard 5-10 had various variants, one being with an Ultra-Linear OP Transformer using 20% taps, and the version without tone controls. In this case, the input signal is applied directly to g1 of the EF86 and a 1M Resistor replaces the volume control, also between g1 & ground. I would suggest keeping the volume control rather.



"Distributed Load .
The arrangement of the output stage for operation under conditions of
distributed load is shown in Fig. 2. The screen-grid resistors R20 and
R21 are disconnected from C14 in Fig. 1 and are connected to the
tappings provided on the primary winding of the output transformer.

The centre tap of the primary winding is connected, as before, to the
reservoir capacitor C15. The dropper resistor R22 in the h.t. line no
longer carries the screen-grid current and so is increased from 1.2kΩ
to 5.6kΩ to maintain the same d.c. conditions in the first two stages
of the amplifier.

The best results are obtained with each half of the primary winding
tapped so that about 20 % of the turns are common to the anode and
screen-grid circuits, when the anode-to-anode loading should be
6.6kΩ."



-F_D
Title: Re: Mars 5-10 stereo - update
Post by: Mars on April 01, 2019, 09:46:37 AM
What would a kit like this retail for? I have been meaning to build a diy tube amplifier for some time now and this looks simple enough for a noob like myself.

Hi JohnnyIC, this kit retails for R9900.
Title: Re: Mars 5-10 stereo - update
Post by: Mars on April 01, 2019, 09:51:18 AM
Yes indeed - Im keen to see how the Mars version differs from the original. The Mars version appears to have no tone control knobs hence my observation

Hi El Sid, I have removed the tone controls and reduced the gain by using a ECC82 phase inverter. Currently I am using 47k and 42k anode load resistors on the ECC82.
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on April 01, 2019, 11:30:42 AM
And have you wound different transformers, or using your standard EL84 PP transofrmers?
Title: Re: Mars 5-10 stereo - update
Post by: Mars on April 02, 2019, 09:49:16 AM
I have simulated the Mullard tone controls in Duncan's Tone Stack Calculator. Is the source impedance correct?

(https://i.imgur.com/cEQP6Ui.png)

(https://i.imgur.com/p1d66r7.png)

(https://i.imgur.com/HNxQPZw.png)

(https://i.imgur.com/493z6Ys.png)

(https://i.imgur.com/3UHcBTK.png)

Next question: I suggest we change the treble pot from 2M to 1M. 2M is hard to find. What would the R4, C3 and C4 values be for such a configuration?
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on April 07, 2019, 09:18:40 AM
Karel,

Ah! Well done!

In the past I noticed that the 5-10 tone controls are somewhat compromised; not a true log-configuration, as is evident from your analyses. Since the input these days is likely to be from low output impedance devices (not a crystal pick-up as was then), one could decrease the tone stack values say some four-fold to make matters more practical. That would take one to 500K potentiometers plus associated components.

You are one ahead of me right now in that my flippen Spice program will not work. Thus I cannot model for different values, but will calculate with pocket calculator to find new values and contact you later, perhaps by PM so that you can put these through your simulation program first. [E.g. the response with controls set at mid-position should be a straight horizontal line, without the 6 dB deviation found by you. (though it seldom is because of making pots true logarithmic in pactice, plus the even greater difficulty of making ganged log pots to track.  Ah well ....)]           
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on April 11, 2019, 02:04:27 PM
Wow ! Lots of interesting facts . Thank you FD , Ampdog and Karel .
Check pm from me Ampdog .
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on April 12, 2019, 06:25:01 AM
With passive tone controls the variable resistors would invariably be of the log taper type. A very preliminary estimate of whether the frequency response will be flat in pot middle position, can be made by checking whether he following relationships are true (referring to Karel's circuits on the left of the graphs):

R1 = 10 x R3
C2 = 10 x C1
C4 = 10 x C3

Unfortunately response at frequency extremities can never be mirror images. In the cut positions (control sliders at R3 and C4 ends) the response will flatten out more than at the opposite sides, i.e. 'boost' will be more than 'cut'.

Symmetrical responses are only possible in feedback circuits e.g. the well-known Baxandall circuit. The responses of channels in stereo would also track better. (Ganged linear pots track better than log ones.) Then R5 there is on the low side; it interferes with the symmetrical operation of the bass pot. It should be at least 6 times the pot value.

 
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on April 15, 2019, 11:40:13 AM
Further to previous post:

Said 'proper' operation of tone controls of log fashion is further only so if nothing else is hung onto the circuit internally or if e.g. the load it feeds into is not high enough. (There is no end to the variations found in practice - goodness knows for what purpose or for whose hearing - lame old topic again!). Most such circuits are best fed from a low input impedance, succeeded by a high load impedance.

Just for general knowledge. And should it be necessary to point out again: At flat settings any properly configured tone control circuit has zero amplitude variation and zero phase shift in its pass band.

[Sorry to seem to always play the policeman; not my purpose. But I wish someone would for a welcome change put up and indicate precisely how tone controls affect (usually meant negatively) sound quality instead of lamely perpetuating ancient folly. The electronics are dead simple.]   
Title: Re: Mars 5-10 stereo - update
Post by: Mars on July 06, 2019, 04:57:57 PM
Schematic

(https://i.imgur.com/5Agj7xI.jpg)

Parts list:

(https://i.imgur.com/zPsnX6t.png)
(https://i.imgur.com/vXT8ZCT.png)
(https://i.imgur.com/E7UOp0W.png)
Title: Re: Mars 5-10 stereo - update
Post by: Den123 on July 06, 2019, 06:10:30 PM
Hey Mars.
Nice clean and simple design. Looks great.
Title: Re: Mars 5-10 stereo - update
Post by: Mars on July 06, 2019, 08:44:06 PM
Thanks to the Stern company :cool:

www.valve-radio.co.uk/literature/stern-mullard-5-10-amplifier-and-control-unit/
 (http://www.valve-radio.co.uk/literature/stern-mullard-5-10-amplifier-and-control-unit/)

.... with a bit of GIMP editing. The numbering follows the original Mullard circuit... so credit is due there too!

www.r-type.org/articles/art-003e.htm
 (http://www.r-type.org/articles/art-003e.htm)
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on July 07, 2019, 09:19:03 PM
Wow Karel ! That valve-radio.co.uk is a super interesting site . Amazing seeing all the vintage equipment .

Thank you kindly Sir .

:)
Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 07, 2019, 06:47:52 AM
Hi All,

I checked with admin, and got an OK to post an update on the kit on the Valves page.

I have a client building a Mars 5-10 preamp and a power amplifier combo on separate chassis. The amplifier is as above, but the pre-amplifier departs from Mullard design. More details later. Here is some photos of the build:

Raw chassis from engineers:

(https://i.imgur.com/bZ4dZIp.jpg)

Customer spay job:

(https://i.imgur.com/LemeQ18.jpg)

Tube layout:

(https://i.imgur.com/VmTtj78.jpg)

Transformers mounted:

(https://i.imgur.com/8lI0NDa.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on October 07, 2019, 11:26:25 PM
Looks Kool Karel . I like the Hammerite finish . The gold on the transformers look a bit gaudy , perhaps a layer of thinned bitumen would knock it back slightly ?

Interesting layout , looks good and will show off all the valves well , looks logical for wiring too .

It's difficult to read the purple text , but from what I can make out :

1. EZ81 rectifier
2. EF86 input voltage gain
3. ECC83 driver
4. EL84 PP output

Well ... I guess that is what a 5 10 is .   :giggle:

Interested to hear about the pre you're planning .
Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 18, 2019, 08:42:28 AM
Next is the pre-amplifier design. My favourite dual pots (ALPS 27) only goes up to 500K audio taper, so I am sticking to tone controls designs that uses 500kA or lower.

Cruising over to an excellent site ( http://www.angelfire.com/electronic/funwithtubes/Amp-Tone-A.html (http://www.angelfire.com/electronic/funwithtubes/Amp-Tone-A.html) ) I chose this design:

(http://www.angelfire.com/electronic/funwithtubes/images/Amp-Tone-1A-A.gif)

Note that the output termination resistance is 500k ohm, so a triode directly after the tone controls are needed.

The next limitation (if you can call it that) is the volume control. I tend to use 100kA pots for attenuation. This is mostly for the direct input line level signals ( the volume pot is usually after the selector switch).

In this case we are looking at this signal flow:

source -> selector -> triode -> tone controls -> triode -> volume pot -> output to power amplifier.

Depending how we design the second triode circuit, we can use lower value pots if we want to. For this design I am sticking to a 12AU7 dual triode to lower the output impedance of the tone control circuit.

Did I miss anything? Does anybody have qualms about a volume pot directly on a preamplifier output?



Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 18, 2019, 09:15:46 AM
proposed tone control and output configuration:

(https://i.imgur.com/kSNEHbh.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on October 18, 2019, 05:52:05 PM
Thanks Karel! I was about to scratch around for some tone circuit ideas for my guitar amp (too dumb/lazy to work it out for myself) so you have saved me some PT
Title: Re: Mars 5-10 stereo - update
Post by: handsome on October 18, 2019, 07:06:53 PM
Volume pot on the output means variable and potentially very high output resistance (25K max in this case). You ideally want a low output impedance to avoid high frequency roll off from cable capacitance and your power amp input capacitance. Why not just use the Alps 500K pot as the output of your tone control? That way your ECC82 will give you a steady ~10k output impedance Its still too high for a modern preamp, you ideally want less than 1k....

Sent from my SM-A750F using Tapatalk

Title: Re: Mars 5-10 stereo - update
Post by: El Sid on October 18, 2019, 08:06:34 PM
Volume pot on the output means variable and potentially very high output resistance (25K max in this case). You ideally want a low output impedance to avoid high frequency roll off from cable capacitance and your power amp input capacitance. Why not just use the Alps 500K pot as the output of your tone control? That way your ECC82 will give you a steady ~10k output impedance Its still too high for a modern preamp, you ideally want less than 1k....

Sent from my SM-A750F using Tapatalk

I.e. replace the 470k resistor with the Alps pot?
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on October 18, 2019, 08:24:18 PM
Volume pot on the output means variable and potentially very high output resistance (25K max in this case). You ideally want a low output impedance to avoid high frequency roll off from cable capacitance and your power amp input capacitance. Why not just use the Alps 500K pot as the output of your tone control? That way your ECC82 will give you a steady ~10k output impedance Its still too high for a modern preamp, you ideally want less than 1k....

Sent from my SM-A750F using Tapatalk

Would a 10 or 20k pot on the output improve things ?
Title: Re: Mars 5-10 stereo - update
Post by: handsome on October 18, 2019, 08:53:39 PM
Yip 500k pot in the 470k's place

A 10 or 20k pot will give you a lower output impedance but would be harder for the ECC82 to drive (more distortion).

Sent from my SM-A750F using Tapatalk

Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 21, 2019, 02:47:21 PM
Like this, handsome?

(https://i.imgur.com/UelUtzG.jpg)

(https://i.imgur.com/yEm1d8L.png)
Title: Re: Mars 5-10 stereo - update
Post by: handsome on October 21, 2019, 04:33:28 PM
Connect the output cap to the junction of the 150R and the 12k resistors. Then your 150R will do double duty: bias the cathode follower and act as a cathode stopper. Good practise to use a grid stoppers as well. Output cap is a bit small for todays lowish input impedances (10k)- use 4.7uF
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on October 22, 2019, 12:41:30 AM
Handsome,

Er - not to nit-pick - neither with you nor this time of night!

But I took it this pre-amp/tone-circuit was specifically for the 5-10 version also shown in this thread (although two volume controls then counters that idea). In that case the output 'loading' is the EF86 grid + 1 meg grid resistor, not any unknown power amplifier ??

Karel,

Did you mean this to be a universal pre-amp/tone control circuit?
Title: Re: Mars 5-10 stereo - update
Post by: handsome on October 22, 2019, 09:39:59 AM
You are correct boss - if just for the 5-10's 1Meg input 220nF is fine.
Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 22, 2019, 10:56:17 AM
output cap moved to 150 ohm/12k ohm junction

grid stoppers added

output coupling capacitor enlarged to 4.7uF

(https://i.imgur.com/q3f3xyI.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on October 22, 2019, 03:03:16 PM
Karel,

There seems then to have been some misunderstanding here. If the above control unit is specifically intended for your 5-10 as per your post #19 schematic, then some simplification can take place. With due respect to Handsome, the coupling cap can revert back to as low as 0,1F, particularly as the advised 4,7F would have to have been an electrolytic. [With an EF86 grid/earth resistor of 100K (the pot) and a 0,1F coupling cap the l.f. pole (-3dB point) will be at 16 Hz; not much on records that low. And for home theatre a totally different set-up will be needed. Should you be more comfortable with some higher value, then fine!]

In that case then the volume control in post #36 may fall away, connecting the 0,1F there directly to the treble pot slider. (No further resistor to earth is required, as earthing is achieved through the bass adjustment chain. The less the loading at that point the more ideal the tone controls will function).

This my take; Apology for apparently buggering you around!
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on October 22, 2019, 10:16:22 PM
It's really fun watching the evolution of this tone control circuit . I'm going to have to get hold of tubecad .

As a really wild guess (cause I really don't know for sure) , I googled "baxandall tone control circuit" , and because I enjoy looking at the angelfire.com studies , decided to check out that page of the myriad images that did appear in the search .

http://www.angelfire.com/electronic/funwithtubes/Amp-Tone-A.html (http://www.angelfire.com/electronic/funwithtubes/Amp-Tone-A.html)

Title: Re: Mars 5-10 stereo - update
Post by: Mars on October 24, 2019, 12:16:56 PM
Ampdog

This preamp should be coupled to a 5-10 power amplifier with no volume control. That way two mono blocks can be controlled with the preamp.

The design dilemma for me is how to match the incoming line level sources (expecting a 100kA, or less,  pot) with the phono stage (expecting a 470k load).
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on October 26, 2019, 01:33:29 PM
Karel,
The value of the input resistor of the first valve is entirely irrelevant. It is there to simply provide a path to connect the input grid to earth.You are feeding a vacuum tube control grid with signal volts. Such a resistor could be 10 ohms but that is a little for most input devices to handle.  It may be up to 10 meg - often used instead of a cathode bias resistor (also often used to provide lower distortion than a cathode bias resistor - but the correct use requires discussion.)

Point here is that the input grid impedance should be high enough for the input device to handle. (Different are the requirements if one drives the tone control input directly. Passive tone controls have substantial variation with frequency, and the ideal there is to drive them with a low impedance of constant value with frequency.

Etc.!
Title: Re: Mars 5-10 stereo - update
Post by: Mars on November 08, 2019, 12:18:35 PM
I've had an request to manufacture an upgrade OPT for the Mars 5-1. I opted for a C-core design and here is the result:

On the photo left you see the regular 15 watt 76mm style OPT. On the right is the larger C-core design.

(https://i.imgur.com/XvIaM5D.jpg)

Title: Re: Mars 5-10 stereo - update
Post by: El Sid on November 08, 2019, 01:09:19 PM
Karel - the usual dof question from me. Why the airgap?
Title: Re: Mars 5-10 stereo - update
Post by: Mars on November 08, 2019, 03:02:54 PM
That was before clamping the cores. Here it is after clamping and varnish:

(https://i.imgur.com/Ev0l4F4.jpg)

(https://i.imgur.com/h27lsnk.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on November 08, 2019, 05:06:20 PM
 :ROFLMAO: like I said dof question
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on November 16, 2019, 08:49:18 PM
Another dof question: is the varnish applied externally, or is the whole transformer dipped?
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on November 17, 2019, 12:19:57 AM
Another dof question: is the varnish applied externally, or is the whole transformer dipped?

If I recall correctly , Karel dips the entire transformer in varnish and then hangs it out to dry .
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on November 17, 2019, 07:14:49 PM
If I may barge/chirp in with my way of applying varnish:

I have the privilege of access to a vacuum pot, courtesy of another member. Thus I am capable of vacuum impregnation - enough suitable varnish to cover the transformer (plus some; some varnish is going to be 'sucked' into the windings!)

Where not available, one can leave the transformer in enough varnish overnight, or perhaps better, apply lavishly with a brush on both ends of the winding. This can be time consuming; don't do both sides simultaneously or the bottom side varnish will flow out all over the show.  Again one leaves until varnish on a particular side has at least gelled. One may also heat up the transformer (BEFORE application!) to drive out some air plus causing the varnish to gel quicker. But disadvantage: If gelling rapidly less varnish will flow in/impregnate.

So in absence of a vacuuming facility I would take the time-consuming route (which I did in the past) of applying varnish, 'thrusting' into layers (if layer=-wound, which I still prefer) with a fairly stiff brush, leaving varnish side up for gravitational ingress, varnish some more, etc. until there is fair certainty that penetration was maximal or has gelled, then flip side etc. etc.  Patience - - but a reasonable degree of penetration can be reached by taking time.

This my way-of-doing-things.  Karel?
Title: Re: Mars 5-10 stereo - update
Post by: El Sid on November 17, 2019, 07:21:34 PM
Another dof question: is the varnish applied externally, or is the whole transformer dipped?

Back off dude! I ask the dof questions around here....
Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 27, 2020, 06:33:11 PM
Hi All

The Mars 5-10 rides again. I have a kit amplifier a client built on the table for fine tuning. I was requested to look at the "hum" that was louder in one speaker than the other, and it moves to the other speaker when the amp is moved!

First I combed through the circuit and cleaned up bad solder joints. Next I added a virtual earth to the filament secondary, and lifted it to 30V DC above ground. That fixed the humm/buzz.

I recently built an amp with EL84 push-pull with a dual choke PSU. It did not have a tube rectifier, but did have an incredible bass response. Thinking back about how this circuit differed from the 5-10 I could list a few points that might or might not have improved bass response:


So I started doing mods and listened to the sound, especially the bass response after each mod. I was hoping one of the interventions alone would shift the bass response to a more prominent position in the sound spectrum.

The easiest was to insert a 8 Henry choke between the EL84 anode and screen grid PSU nodes. That did improve bass and the sound in general.

Next I increased coupling capacitors to 0.22uF 400V. Bass improved, so another good idea.

Then I did the cathode bypass capacitors. 1000uF 16V all around. Bass improved again.

I had a look at TubeCAD calculator. I typed in the circuit values (47k anode load resistors/22k cathode long tail/260V supply). TubeCAD did not like this arrangement. It suggested I rather try 15k and 18k load resistors and a 9k long tail. I used a 10k long tail and 33k//27k and 33k//39k anode resistors. Bass does sound better? Or am my ears lying? Let me leave it on for a while. Voltages measures safe.

Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on January 28, 2020, 10:13:56 AM
Interesting post, thanks.

I have wondered for a while if 0.1uF is too low a value for the coupling caps. Maybe bigger caps were too big and/or expensive back in the late 50s? Perhaps that affected the size of the cathode bypass caps too?
Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 28, 2020, 04:22:05 PM
Thanks for the reply Qualityten.

If I look at the maths, I get these results: (The load resistor after the phase inverter coupling capacitor is 470k Ohm in this circuit.)

That means that (using Xc=1/(2*PI*f*C)   f   for 0.1uF is 3.4Hz and for 0.22uF  1.5Hz

Not a big difference on paper.

So why did the ancients use 0.1uF caps rather than 0.22uF or 0.33uF? My guess was that the speakers could not reflect those low notes we hear in modern music. Another issue would be the stability of the circuit with large amounts of negative feedback (as in the original 5-10 circuit). Maybe someone else can comment?

Ampdog has mentioned before how large cathode bypass capacitors can have the effect of fixed bias; thus less bias fluctuation during music playback.
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on January 29, 2020, 01:35:42 AM
Thanks Karel. I need to go back in a time machine and do an elec eng course at a time when I still remember my school science and maths. What is Xc in your formula, and why are the load resistors increased from 4k7 to 470k in your circuit?

Your last point is intriguing too!
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on January 29, 2020, 05:57:46 AM
Thanks Karel. I need to go back in a time machine and do an elec eng course at a time when I still remember my school science and maths. What is Xc in your formula, and why are the load resistors increased from 4k7 to 470k in your circuit?

Your last point is intriguing too!

Look here : http://sim.okawa-denshi.jp/en/CRhikeisan.htm (http://sim.okawa-denshi.jp/en/CRhikeisan.htm)
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on January 29, 2020, 07:38:44 AM
My turn to thank you, fredeb.
Title: Re: Mars 5-10 stereo - update
Post by: Jacobus on January 29, 2020, 07:48:43 AM

So why did the ancients use 0.1uF caps rather than 0.22uF or 0.33uF? My guess was that the speakers could not reflect those low notes we hear in modern music. Another issue would be the stability of the circuit with large amounts of negative feedback (as in the original 5-10 circuit). Maybe someone else can comment?


Reading through this give a possible reason.

https://www.aikenamps.com/index.php/what-is-blocking-distortion
Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 29, 2020, 02:29:03 PM
Hi Qualitten

The resistor value you need to use is the grid leak for the following EL84.

(http://sim.okawa-denshi.jp/images/CRHiPass.png)

R=470k
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on January 29, 2020, 10:45:57 PM
Hi Qualitten

The resistor value you need to use is the grid leak for the following EL84.

(http://sim.okawa-denshi.jp/images/CRHiPass.png)

R=470k

Hey Karel , I'm interested to know whether 470k is the max permissible value allowed for that resistor in an EL84 amp ? As this this value already appears to be high . And just guessing , but I suppose the higher this value is , the easier it is to drive the EL84 ?
Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 30, 2020, 07:14:23 PM
fredeb, the data pages gives 1M as max value for cathode bias. Yes, the higher the grid leak resistor value the easier the load on the phase inverter. But with a medium mu tube like the ECC82, lower values could be tried.

I have also found that the lower the grid leak resistor value, the better the output tube sounds; with transformer drive being the best sounding. So I generally go as low as I can.

I tried 220k value for grid lead in this amplifier, and the results seems to be better. So I am staying with it for the moment.
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on January 30, 2020, 07:46:14 PM
fredeb, the data pages gives 1M as max value for cathode bias. Yes, the higher the grid leak resistor value the easier the load on the phase inverter. But with a medium mu tube like the ECC82, lower values could be tried.

I have also found that the lower the grid leak resistor value, the better the output tube sounds; with transformer drive being the best sounding. So I generally go as low as I can.

I tried 220k value for grid lead in this amplifier, and the results seems to be better. So I am staying with it for the moment.

Thanks for the explanation Karel , appreciated man !  :thumbs:
Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 31, 2020, 05:38:04 PM
After a bit of listening, here is an update of where the tweaking has led me:

The pentode mode, for all it's bass, had me up against the walls with the piercing high frequencies. So back to ultra-linear. That means the screen choke goes out and a 22uF filter cap and 1 Henry  20 ohm DCR choke is added after the GZ34, before the 100uF 450V filter cap. I checked with PSUD2 and no red lights flashed regarding the 100uF cap.

The sound is much better. Next I looked at the Zobel network parallel to the EF86 anode load resistor.  The load resistor is 120k ohm. I paralleled a 1nF capacitor to roll off high frequencies above 1kHz.

So to sum up:

(https://i.imgur.com/5Agj7xI.jpg)

Title: Re: Mars 5-10 stereo - update
Post by: Mars on January 31, 2020, 05:45:24 PM
Another dof question: is the varnish applied externally, or is the whole transformer dipped?

The whole transformer gets dipped (I leave it a hour or two to soak properly) and hung up in the air to dry.
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on January 31, 2020, 10:37:35 PM
Just curious: Karel, can you say any more about why reducing the grid leak resistors, R12 and R13 down to 220k should improve the sound?  Bearing in mind that the original circuit has 820k there.

Thanks again for sharing your findings.
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 01, 2020, 07:08:57 PM
Chipping in here with random remarks, if I may:

1.. For grid resistors for EL84 (and most other applications) 470K tends toward maximum. In early days specs stated a high 820K max allowed for EL84, but particularly Leak learned an expensive lesson that that was too high (don't know which bright brain at Mullard worked that out). Run-away* is particularly possible with close-spaced grid-cathode construction and high temperature as in the EL84, and is a merciless 'lesson'.  (Leak had to withdraw his early amps and replace.) Plus what with spread in characteristics (should be a constant at the back of the engineer's mind!) . . . beware! Yes, this limits the driver valve's anode resistors, but such is life!  (Bear in mind that for fixed bias, where compensation from cathode resistors are absent, this limit is mostly 100K - you disregard this at your peril :(  ) That often complicates design of high output final stages.)

2. I would not call C13-R22 a Zobel in this case, rather a phase corrective measure plus limiting open loop h.f. gain in order to stabilise NFB. Sort of similar operation but different idea.  (And here - oh boy!  Kindly pardon my chagrin. Yes - proportion those to get the NFB stable - - and decrease loop gain to the extent that very audible distortion results at h.f., because of decreasing NFB!  How many amplifiers gaily starts cutting loop gain upwards of 1 KHz - and then wonders why listener fatigue sets in. Basics: One is supposed to have full NFB active over the whole audible range!!  NFB and its 'rules' are supposed to make a good amplifier better - not tinker up a poor design to something-  er - mm.)

But there I go again ,,,
Sorry  (only I get upset when results are pursued at the cost of basics  >:(
___________________________________________________________________
*G1 runaway: Stray electrons land on G1 rendering it slightly positive, more electrons land there unable to leak away quickly enough, etc.  Go internet.
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on February 01, 2020, 08:30:46 PM
Thanks for that Ampdog , your knowledge is always appreciated .

Sent from my SM-G935F using Tapatalk

Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 01, 2020, 09:00:20 PM
Chipping in here with random remarks, if I may:

1.. For grid resistors for EL84 (and most other applications) 470K tends toward maximum. In early days specs stated a high 820K max allowed for EL84, but particularly Leak learned an expensive lesson that that was too high (don't know which bright brain at Mullard worked that out). Run-away* is particularly possible with close-spaced grid-cathode construction and high temperature as in the EL84, and is a merciless 'lesson'.  (Leak had to withdraw his early amps and replace.)

Ampdog, thanks for your comments.  I see that the Leak TL12 plus has 470K there, but the Leak Stereo 20 uses 1M.  Am I missing something?
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 02, 2020, 09:36:37 AM
You are missing the previous post!  :)

That was where I explained the unfortunate use of 1M grid resistors for EL84s! In initial data sheets Rg1 max, were stated as 680K, and in other data sheets even as 1M as said. Leak had to recall numbers of 20/20s to retrofit with 470K resistors, (Here somewhat of a slip in the great man's logic: one never uses a component value at its maximum specified value - what about spreads etc?  Harold?? )

In the subsequent Stereo 12+ the lesson was apparently learnt, the grid resistors rapidly dropped to 470K, which also became the recommend value for subsequent grid resistors. (All then concerning EL84 grid resistors.)
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 03, 2020, 12:06:37 AM
Thanks Ampdog!  I did read your earlier post, but see that the Leak Stereo 20 schematics on the web still have 1M there, including the authoritative site here:  http://44bx.com/leak/Leak/Circuits/Stereo20.gif
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 03, 2020, 03:50:06 AM
Qualityten,

Yes, that is the official Stereo 20 schematic. It does not appear that a different schematic was issued after the 'recall' early in the 1960. In a brochure about Leak called "Firsts in High Fidelity"* by Stephen Spicer (p.154) the recall is briefly mentioned.

I am uncertain where I first learned about it. (I was then employed by the C.S.I.R. and several periodicals were in circulation for compulsory reading.) When H.J. Leak visited Cape Town 1965/66, a couple of us students  attended a demo talk and had tea with him afterwards,. but I do not think this was discussed then - not sure when the Stereo 20 was produced.

It appears no exception that maximum specifications for valves are often exceeded, particularly in guitar amps  (consequences are suffered afterwards.)
________________________________________
* Printed by Audio Amateur Press, U.S.A.
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 04, 2020, 01:02:00 PM
Thanks Ampdog.  Your expertise and recollections are such an asset to this forum. 

Off topic: Would you care to make a post about your recollections of meeting H. J. Leak?  It would be particularly interesting to hear of your assessment of him as an engineer, as distinct from being an entrepreneur.  I have not managed to get a copy of Firsts in High Fidelity yet.  I see there is one on offer, secondhand, for 1,300 and one sold recently on ebay for around 250!
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 04, 2020, 11:07:04 PM
As a p.s. to my last post, I read the Hi-Fi World DIY Supplement no 33 for Dec 1997, which happend to have a letter asking about modifications to the Leak Stereo 20.  The circuit diagram published with the reply by Chris Found has 470K resistors for R12 and R13.  I don't know the source of that version of the diagram, but the change was clearly standard by then.  With apologies to Karel for this Leak digression in his thread.
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 05, 2020, 01:01:54 AM
O Gaats!!

Absolute apologie, Karel! Daar gaan ek alweer oor iets wat ek opgemerk het. Ek skuld 'n gratis middagete as U weer hier rond is!

Back to the question in post #62 and topic.  My apologies for taking off on a tangent!  (Next time Karel, alert  the moderators!)

(Ek vertel graag iets oor H.J.Leak in 'n ander 'thread'.)
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 05, 2020, 02:01:01 AM
At least the digression was related to Karel's changes to the grid leak (Leak?!) resistors from 470k down to 220k.
Title: Re: Mars 5-10 stereo - update
Post by: MusicMan_ZA on February 05, 2020, 06:56:01 PM
O Gaats!

(Ek vertel graag iets oor H.J.Leak in 'n ander 'thread'.)

Asseblief!
Title: Re: Mars 5-10 stereo - update
Post by: Mars on February 06, 2020, 01:17:46 PM
Here is some screenshots from the TubeCAD program that shows the calculations for 470k grid leak resistors versus 220k grid leaks:

(https://i.imgur.com/pACSpSS.png)

470k grid leak

(https://i.imgur.com/RGTXJcJ.png)

220k grid leak

(https://i.imgur.com/BpTLhPZ.png)

Comments welcome...
Title: Re: Mars 5-10 stereo - update
Post by: handsome on February 06, 2020, 02:07:27 PM
Neither option looks very good....you only have 11V of negative swing. So you could drive an EL84 but just and only if it is biased to that - but you will be at your limits so more distortion than necessary. A higher HT, a negative supply or less current will help and let you use a bigger anode resistors; ~17k is a low load for a 12AU7.

Sent from my SM-A750F using Tapatalk

Title: Re: Mars 5-10 stereo - update
Post by: Mars on February 06, 2020, 05:20:38 PM
Thanks for the comment handsome...
Now I understand why Ampdog leans toward ECC88 (from comment somewhere, if memory serves me well :thinking:) Much better figures at 280V B+

I aimed for a 50% voltage drop across the triode, but 11k is as low as the anode resistor is allowed by the program.

ECC88

(https://i.imgur.com/CDgLSOG.png)
Title: Re: Mars 5-10 stereo - update
Post by: Mars on February 07, 2020, 03:12:04 PM
Here is a photo of the amp under discussion

(https://i.imgur.com/du81F3H.jpg)
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on February 07, 2020, 05:04:51 PM
That's a good looking amp Karel , certainly looks like it should have a bit of vooma for an EL84 PP .

How do things look when you double or triple the ecc82 anode resistor ? Do you mind doing that in the software so we can see the difference please .


That said , I'm definitely going to purchase tubecad , seems it's only $10
Title: Re: Mars 5-10 stereo - update
Post by: Mars on February 12, 2020, 11:34:31 AM
As requested:

(https://i.imgur.com/NoFX0ZT.png)

and report

(https://i.imgur.com/0gPsdGT.png)

As you can see, the 12AU7 does not perform well in this circuit. The easiest way to make this work is with a increased B+ or a negative voltage applied to the long tail resistor. 150V seems to be the ideal.

(https://i.imgur.com/G8bSl49.png)

I think I can find a way of adding a negative 150V supply from the PSU without too much sweat. But using an ECC88 makes all problems go away, except you cannot use only tubes robbed from a guitar amplifier :sd:

Or, stick with guitar amp tubes and deal with the high sensitivity in another way:

12AX7 with original values as in Mullard 5-10

(https://i.imgur.com/X6F7J5i.png)
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 12, 2020, 06:11:36 PM
Qualityten,

Yes, that is the official Stereo 20 schematic. It does not appear that a different schematic was issued after the 'recall' early in the 1960. In a brochure about Leak called "Firsts in High Fidelity"* by Stephen Spicer (p.154) the recall is briefly mentioned.

I am uncertain where I first learned about it. (I was then employed by the C.S.I.R. and several periodicals were in circulation for compulsory reading.) When H.J. Leak visited Cape Town 1965/66, a couple of us students  attended a demo talk and had tea with him afterwards,. but I do not think this was discussed then - not sure when the Stereo 20 was produced.

It appears no exception that maximum specifications for valves are often exceeded, particularly in guitar amps  (consequences are suffered afterwards.)

Picking up this point from Ampdog, I raised the question of when Leak changed the value of the grid leak resistors in the Stereo 20 on another forum and got this interesting reply:
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on February 12, 2020, 08:54:15 PM
Thanks for indulging me Karel , it appears that many values would need to change . I almost purchased tubecad , but noticed the non 64-bit operating system warning , I suppose it wouldn't cost too much to setup a tubecad dedicated machine running 32-bit XP .

BTW ... 6N1P , could be an affordable replacement for ECC88 now that 6N23P and the -EV version are becoming so expensive .
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 13, 2020, 07:34:22 AM
Basics again (OT, but at least Karel gets the exposure :)!

In voltage amplifiers one stays away from load resistors close to the valve value of rp .  Rl = rp is the condition for maximum power transfer (not required in voltage amplifiers), but also for maximum distortion. Thus for say 12AU7, rather higher than 15K, for pentodes either <220K or >1M (the latter becoming impractical for other reasons).  For ECC83, rather >39K etc. (But as always with electronics, not quite that simple: rp under specific operating conditions; rp changes significantly with current).

Qualityten:
Thanks. Yes, easy to comment after the fact, but the man who originally reckoned that >470K is safe EL84 grid resistors with that grid construction (evident from the high gm):  :o

But again: Back to topic (Karel is most indulgent!)
Title: Re: Mars 5-10 stereo - update
Post by: Mars on February 18, 2020, 10:17:30 AM
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 18, 2020, 02:30:32 PM
The excess gain is in the first stage, with the EF86, so I don't think the different taps would affect that issue.  However, 20% screen taps are recommended as optimal, rather than the more usual 43%.  I don't know enough theory to know why!
Title: Re: Mars 5-10 stereo - update
Post by: handsome on February 18, 2020, 08:17:34 PM
To reduce gain: strap the ef86 as a triode and unbypass its cathode resistor
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 19, 2020, 03:38:17 PM
I have had good results with triode strapping EF86s in my Verdik Qualityten 5-10 amps, which otherwise follow the classic Mullard circuit exactly. 

I followed the guidance in this thread on the UK Vintage Radio Repair and Restoration forum. https://www.vintage-radio.net/forum/showthread.php?t=99399. See posts #7 and #16 for circuits.
Title: Re: Mars 5-10 stereo - update
Post by: Ampdog on February 19, 2020, 07:40:05 PM
To reduce gain: strap the ef86 as a triode and unbypass its cathode resistor

. . . . and consequently reduce the  feedback resistor accordingly - otherwise the above steps will merely be countered by decreasing the NFB, without causing a notable reduction in gain.
Title: Re: Mars 5-10 stereo - update
Post by: handsome on February 19, 2020, 09:02:53 PM
Whoops! forgot about the feedback <blush> you are, as ever, quite correct, Boss - judicious diddling with feedback resistors, compensation networks and inviting a square wave to the party would be in order then
Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on February 22, 2020, 12:28:07 PM
It so happens that the same question was asked in a post on Audiokarma.  A knowledgeable member there commented as below.  This helps explain handsome's advice to remove the cathode bypass cap:

'The cathode bypass cap will increase open loop gain. Amps that apply a lot of feedback may need more open loop gain for a given input sensitivity. Id try it without cathode bypass. See if you can live with the input sensitivity given the amount of feedback you want to apply.  Also, if you triode strap, you may need to rebias the EF86, meaning you may need to change the value of the cathode resistor to give you the bias point you want. Since the EF86 is direct coupled to the phase inverter, you do want the EF86 plate voltage the same as in the original circuit where it was pentode wired.'
Title: Re: Mars 5-10 stereo - update
Post by: handsome on February 22, 2020, 03:32:40 PM
The open loop gain of an amplifier is its 'inherent' gain - the product of the amplification of all the stages  - without any feedback applied
The closed loop gain of an amplifier is its gain after negative feedback has been applied - it is: the open loop gain (in dB) - feedback (in dB) = closed loop gain (in dB)
The closed loop gain of an amplifier is defined by it's feedback resistors (as long as it's open loop gain is significantly larger than the desired closed loop gain)
The main advantage of feedback is that the gain is now dependent on the resistors and as long as they don't drift in value (they shouldn't) the gain will be stable even as the valves age provided the open loop gain remains significantly larger than the closed loop gain. So your valves may be mismatched between channels or age at different rates your gain will remain the same.
So reducing, as I erroneously suggested, the gain of any stages in an amplifier using feedback will only reduce the amount of feedback, not the gain of the amplifier.

So you have an amplifier with 60dB of open loop gain (x1000) and 1% distortion, you use a combination of feedback resistors to give the amplifier a closed loop gain of 20dB (x10) you have then applied 40dB (x100) of feedback and can expect your distortion to be 0.01% (/100). Sounds good ....but.... lots of feedback brings with it the possibility of making your amplifier unstable or turning it into a full time oscillator - and this is the reason for the compensation networks you see in amplifiers they aim to manipulate the phase and frequency response of the amplifier in order to keep it stable for the amount of feedback being used (more complex discussion needed).

Negative feedback has other advantages: generally speaking negative feedback will reduce distortion, increase bandwidth, decrease output impedance and increase input impedance depending on how it is derived and applied (more complex discussion needed).
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on February 23, 2020, 03:55:47 AM
Thanks for setting this out so comprehensively handsome . It helps a lot with ones undersanding thereof .

Negative feedback has other advantages: generally speaking negative feedback will reduce distortion, increase bandwidth, decrease output impedance and increase input impedance depending on how it is derived and applied (more complex discussion needed).

I'd love to hear more .
Title: Re: Mars 5-10 stereo - update
Post by: Mars on July 28, 2020, 12:30:19 PM
Hi All

We are picking up speed again with projects, after the inconveniences of the last few months.
Here is a version of the 5-10 with a bluetooth input channel.

(https://i.imgur.com/ggo934q.jpg)

(https://i.imgur.com/HrnwO7f.jpg)

(https://i.imgur.com/1WfRFPI.jpg)

The output transformer has been made with three bifiliar secondaries in parallel for better coupling and less phase shift (that the usual two secondaries).

(https://i.imgur.com/Uzsx89m.jpg)

(https://i.imgur.com/HNsVS8H.jpg)

Title: Re: Mars 5-10 stereo - update
Post by: El Sid on July 28, 2020, 01:14:26 PM
Nice Karel! I like the EF86s...
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on July 28, 2020, 01:56:54 PM
Amazing work Karel ! These amps must sound amazing .
Title: Re: Mars 5-10 stereo - update
Post by: Family_Dog on July 28, 2020, 02:44:54 PM
The 5-10 is a decent all-rounder, it does sound sweet.


-F_D
Title: Re: Mars 5-10 stereo - update
Post by: LAV on July 28, 2020, 05:18:21 PM
Hi All

We are picking up speed again with projects, after the inconveniences of the last few months.
Here is a version of the 5-10 with a bluetooth input channel.

(https://i.imgur.com/ggo934q.jpg)

(https://i.imgur.com/HrnwO7f.jpg)

(https://i.imgur.com/1WfRFPI.jpg)

The output transformer has been made with three bifiliar secondaries in parallel for better coupling and less phase shift (that the usual two secondaries).

(https://i.imgur.com/Uzsx89m.jpg)

(https://i.imgur.com/HNsVS8H.jpg)

BAAAIE mooi!
Title: Re: Mars 5-10 stereo - update
Post by: Mars on August 24, 2020, 12:19:06 PM
Hi All
I have built another variant on the circuits discussed, with good results.

To start at the beginning: the PSUD2 predicted a DC voltage of 340V B+ for the power transformer I built and the current ratings I expected. This was for 330R cathode resistors on all 4 EL84's.

Since that was marginally too high, I though it would be a great idea to try out the Blumlein garter circuit again for perfect DC balance on the OPT's.

I came up with an elegant layout that worked tremendously.

(https://i.imgur.com/n1BaQNi.jpg)

But... the B+ turned out to be 310V; exactly right for regular Mullard 5-10 circuit. Next problem: PSUD2 told me that I exceeded forward current of 0.75A limit. Increasing the supply resistance did not help, neither changing filter capacitor values. Only one thing; run the EL84's cooler. Like with a 390R cathode resistor.

The idea to change 8 cathode resistors on the Blumlein circuit was to much to ask, with the client already on their way to collect.

So the Garter circuit was scrapped (to done on another day) and the 4 x 390R's fitted.

(https://i.imgur.com/K7xLivM.jpg)

Next detail: I tried a split load phase inverter to reduce the gain. That means I used only one half of the ECC82. I configured the two phase inverters in such a way that Left and Right channel uses a different half of the double triode. That way you can just swop your two ECC82's when one half of the tube becomes weak or broken.

Look at the two sockets left and right below the row of 4 EL84's and one GZ34

(https://i.imgur.com/mW0W29c.jpg)

I had to adjust the power supply feed resistor to the EF86's for a better DC spread across the 33k cathode and anode load resistors of the PI.

The outcome was an amplifier with a much better sensitivity than previous and a wonderful clear sound.


Title: Re: Mars 5-10 stereo - update
Post by: Qualityten on September 02, 2020, 06:02:07 PM
Thanks, as always, for sharing of your work.  Very enjoyable and instructive!
Title: Re: Mars 5-10 stereo - update
Post by: marantz123 on September 04, 2020, 01:22:07 PM
Sire, wonderful work ! I see you have Russians in there (6p14p-EV/B) ...have you ever had a bad one, and how do these compare sonically to the standard EL84?

Being an OTK graded tube, are they already reasonably matched into a specific range, possibly no matching issues with any pair of tubes chosen or have you found them to be far out of each other sometimes ?

Also, any experience with the 6n2p-EV as a sub (par the heater arrangement) for the 12AX7 ? How do they compare ?

Im venturing into Russian territory, hence an opinion would be welcome.
Ta.
Title: Re: Mars 5-10 stereo - update
Post by: fredeb on September 04, 2020, 05:44:57 PM
John Broskie's take on the  Blumlein Garter circuit  :    https://www.tubecad.com/2009/04/blog0163.htm (https://www.tubecad.com/2009/04/blog0163.htm)


Did it work at all Karel ?
Title: Re: Mars 5-10 stereo - update
Post by: Mars on September 07, 2020, 02:34:34 PM
Hi fredeb

It did work. I abandoned the idea when I had to change the cathode resistors to 390 ohm to run the tubes cooler and keep the GZ34 happier.