Author Topic: Sansui 1000A Transformers on chassis  (Read 1862 times)

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Sansui 1000A Transformers on chassis
« on: February 14, 2017, 10:10:47 PM »
I recently purchased ( from John ) a chassis with 2x Sansui 1000A OPT's , 1x Sansui 1000A power transformer with RCA plugs , Binding posts , Bias pots and measuring points and some sockets fitted . Transformers do not have UL taps and power transformer taps are 190V , 32V and 3x 6.3V ; primary taps are 115 or 230V . OPT's are 30W specimens designed for 2x 7591 @ 420v on the anodes . I'm guessing then , that the OPT's are wound for a ~6600K plate to plate impedance , my guestimate coming from the 7591 datasheet specs for UL PP specs .

Can anybody please help me figure out what plate to plate impedance the OPT's are set up for ? Secondary taps are rated at 4 , 8 and 16ohm on the schematic .

Also , I wondering if it is possible to get 30W output from any power tubes in triode or pentode mode without using a voltage multiplier , but using ~280VDC B+ from solid state rectification ?  Could I get that from 2x 6L6's ? Any suggestions and thoughts welcomed . What would you do ? 

7591 datasheet http://tubedata.tigahost.com/tubedata/sheets/168/7/7591.pdf

I'd like to use both triodes of a 6N6P ( ~E182CC ) as drivers for output tubes , and perhaps another 6N6P ( or 6N1P or 6N2P ) for gain and phase-splitting .

Schematic provided for idea of transformer components .



And pics of blank canvas .







“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Offline Qualityten

  • Trade Count: (0)
  • AVForums Member
  • **
  • Posts: 96
Re: Sansui 1000A Transformers on chassis
« Reply #1 on: February 15, 2017, 12:00:12 AM »
Is the issue that you can't get, or don't want to use 7591s as the power tubes?  I have rebuilt a Eico ST40 amp designed for 7591s to use 6L6s, but these were connected in triode mode.  It may not be what you want, but it may give you some ideas.  http://www.audiokarma.org/forums/index.php?threads/eico-st-40-redo-modo.211563/page-2

Offline Ampdog

  • Trade Count: (+1)
  • AVForums Veteran
  • *****
  • Posts: 6,695
Re: Sansui 1000A Transformers on chassis
« Reply #2 on: February 15, 2017, 03:16:12 AM »
Can anybody please help me figure out what plate to plate impedance the OPT's are set up for ? Secondary taps are rated at 4 , 8 and 16ohm on the schematic .

Likely 6600 ohm -  but you can determine that easily by feeding a some 100 volt (50 Hz) into the primary and calculating the primary impedance from the transformer ratio.

Quote
Also , I wondering if it is possible to get 30W output from any power tubes in triode or pentode mode without using a voltage multiplier , but using ~280VDC B+ from solid state rectification ?  Could I get that from 2x 6L6's ? Any suggestions and thoughts welcomed . What would you do ?

Unlikely, as the OPT ratio does not permit a low enough primary impedance (high enough current) to reach 30W. Preferable to go the voltage-doubler power supply route. Any reason why not?

Best triode operation is with higher load impedance (around 8K - 10K), thus high power supply voltage - in the 450V - 500V region. 

Quote
I'd like to use both triodes of a 6N6P ( ~E182CC ) as drivers for output tubes , and perhaps another 6N6P ( or 6N1P or 6N2P ) for gain and phase-splitting .

Something along Williamson line? A very good approach, with the right component values.

Note: The 7591 is somewhat of a 'supped-up' EL84; higher dissipation, output power at higher permitted voltages. In that sense different from a 6L6. The 6L6GC is rather more freely available but requires twice the grid signal voltage - not a problem with the pre-circuit at your determination. (However hardly driveable by the popular 'cathodyne' phase splitter - don't go that route as the higher power Dynacos did.)

Pity about no screen taps. Still, many pentode amplifiers existed with pentode feed, if somewhat more sensitive to output load (loudspeaker) and NFB stability. As said, triode performance would be compromised due to low primary impedance.

What I would do?

Use voltage doubler power supply etc. plus pentode operation and Williamson or Leak type topology, and hope the OPT plus associated circuitry will allow enough NFB to give good fidelity. Has been done.  (You should have enough h.t. to reach 30+W output with cathode-bias.)

Caveat: The E182CC is a good driving twin triode but the spread can be daunting because they were intended for computor use - low 'on' voltage and long no-current life - rather than analogue consistency. I use them but have to select.
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #3 on: February 15, 2017, 04:35:41 AM »
Is the issue that you can't get, or don't want to use 7591s as the power tubes?  I have rebuilt a Eico ST40 amp designed for 7591s to use 6L6s, but these were connected in triode mode.  It may not be what you want, but it may give you some ideas.  http://www.audiokarma.org/forums/index.php?threads/eico-st-40-redo-modo.211563/page-2

I'd like to stay with cheaper , more readily available tubes , I have quite a couple of 6P3S's ( ~6L6GT ) , some EL34's , some KT88 ( these appear to be quite expensive though ) , some GU-50 ( ~LS50 , I have sockets for these too , if I do then use voltage multiplier , output loads of power easily enough , but at very high voltage ) .

Thanks for the link , I'll check it out , perhaps I'll learn something . :)

Likely 6600 ohm -  but you can determine that easily by feeding a some 100 volt (50 Hz) into the primary and calculating the primary impedance from the transformer ratio.


That's how you do it ! I remember watching a youtube video by " Uncle Doug " and his funny dog now that you mention it , just needed a reminder , thank you Ampdog !
 
Unlikely, as the OPT ratio does not permit a low enough primary impedance (high enough current) to reach 30W. Preferable to go the voltage-doubler power supply route. Any reason why not?

I wanted to keep it simpler , if it were possible , but if i have to go the multiplier route - so be it . Did you look at the multiplier in the original circuit Ampdog ? HT linked to bias supply . Is that an " OK " way to go ? Or rather keep them separate ?
 
Best triode operation is with higher load impedance (around 8K - 10K), thus high power supply voltage - in the 450V - 500V region. 

So the voltage multiplier can definitely NOT be avoided then . :)
 
Something along Williamson line? A very good approach, with the right component values.


Yes , exactly , I think 30W from a pair of KT88's or LS50's in triode/pentode mode can be had with relative ease ( with higher HT then ) . 

Note: The 7591 is somewhat of a 'supped-up' EL84; higher dissipation, output power at higher permitted voltages. In that sense different from a 6L6. The 6L6GC is rather more freely available but requires twice the grid signal voltage - not a problem with the pre-circuit at your determination. (However hardly driveable by the popular 'cathodyne' phase splitter - don't go that route as the higher power Dynacos did.)

Pity about no screen taps. Still, many pentode amplifiers existed with pentode feed, if somewhat more sensitive to output load (loudspeaker) and NFB stability. As said, triode performance would be compromised due to low primary impedance.


If the 30W is do-able with 6P3S ( with no screen taps ) , I'd like to use them just because I do have loads of them . If not , I'd prefer to go either EL34 or GU50 route , for economic reasons . For the money I reckon the KT88's can be put to better use elsewhere .

What I would do?

Use voltage doubler power supply etc. plus pentode operation and Williamson or Leak type topology, and hope the OPT plus associated circuitry will allow enough NFB to give good fidelity. Has been done.  (You should have enough h.t. to reach 30+W output with cathode-bias.)

It would be foolish of me to ignore your wisdom and experience , so this is then what I'll do . Do you not think the fixed bias is possible ? Would be good for unmatched valves ?

Caveat: The E182CC is a good driving twin triode but the spread can be daunting because they were intended for computor use - low 'on' voltage and long no-current life - rather than analogue consistency. I use them but have to select.


I have a stash of 16x 6N6P's that have been measured and marked at least , I'll use the balanced/matched triodes pairs for drivers , and unbalanced/matched pairs for input/phase-splitter .

“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Offline Mars

  • Commercial Member
  • Trade Count: (0)
  • ***
  • Posts: 924
  • “The path is the goal.” Mahatma Ghandi
    • Mars Tube Audio
Re: Sansui 1000A Transformers on chassis
« Reply #4 on: February 15, 2017, 05:37:13 AM »
Nice project, fredeb! I am following the thread with interest :2thumbs:

OEM manufacture of Tube Audio
Power and Integrated Amplifiers
Phono Pre-amplifiers
DIY Kits
DIY Workshops

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #5 on: February 15, 2017, 07:08:58 AM »
Nice project, fredeb! I am following the thread with interest :2thumbs:

Thank you Karel . I'm working 5/6 days a week , 12 hours a day again , so progress will be slow . But may as well staart getting an idea of which direction I'll be going .  :D
“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #6 on: February 15, 2017, 09:27:03 PM »
Alrighty , from what I can tell , the 6P3S seems to be identical to a Brimar 6L6G and 6L6GT shown at r-type.org .

6L6G : http://www.r-type.org/exhib/aaa1590.htm



6L6GT : http://www.r-type.org/exhib/aaa0016.htm





Datasheet for Tungsol 6L6G from r-type : http://www.r-type.org/pdfs/6l6g.pdf

Now , interestingly enough , the datasheet depicts these values for Class AB1 in pentode mode ...



..... and these values for Class AB2 in pentode mode ....



What's the difference between the 2 depicted load resistances ? Perhaps a 1,5K resistor in series with each anode ?

I'd love to see the schematic whereby these conclusions were come to .
« Last Edit: February 15, 2017, 09:29:23 PM by fredeb »
“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Offline Ampdog

  • Trade Count: (+1)
  • AVForums Veteran
  • *****
  • Posts: 6,695
Re: Sansui 1000A Transformers on chassis
« Reply #7 on: February 16, 2017, 02:20:24 AM »
Firstly, output power is not simply working into a certain load impedance, but also transferring all of the valve generated power to the load. Putting resistors in serie with the anode(s) may give the 'right' impedance, but the power dissipated in the resistors is heat and does not contribute to the output.

As to the tables, one looks at class AB1 operation. (In class AB2 some input grid power is drawn and this is not generally applicable to quality audio amplifiers.)  The lower power (18W) into a 3K8 load compared to 26,5W into 6K6 ohm is rather abnormal. (Usually output power increases with decrease in load impedance, as reflected in the higher peak current drawn, all other things being equal. I did not go to plotting graphical analysis for a check at this time; let us leave it at that.)

Basic Ohms Law arithmetic will show that to get 30W into a 6600 ohm load in push-pull, some at least 430V h.t. is required (taking an anode saturation dip of 80V) - this is irrespective of the valve used.  One is therefore limited by Ohms Law regarding maximum output power available.

[Also keep in mind that valve data sheets give power available at the valve anode(s).  The efficiency of the OPT (at best some 90%) comes afterwards, thus a tabled output of 30W will give only 27W to the loudspeaker.]
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #8 on: February 16, 2017, 05:00:32 AM »
Firstly, output power is not simply working into a certain load impedance, but also transferring all of the valve generated power to the load. Putting resistors in serie with the anode(s) may give the 'right' impedance, but the power dissipated in the resistors is heat and does not contribute to the output.

As to the tables, one looks at class AB1 operation. (In class AB2 some input grid power is drawn and this is not generally applicable to quality audio amplifiers.)  The lower power (18W) into a 3K8 load compared to 26,5W into 6K6 ohm is rather abnormal. (Usually output power increases with decrease in load impedance, as reflected in the higher peak current drawn, all other things being equal. I did not go to plotting graphical analysis for a check at this time; let us leave it at that.)

Damn ! It's probably a typo in the datasheet ( Class AB1 section , AB2 is the other way round ) . I was wondering about the quality of Class AB2 and if it is worth pursuing .

Should I be be aiming at Class A PP ? Can one have the choice of Class A or AB1 in the same circuit depending on bias ?



Basic Ohms Law arithmetic will show that to get 30W into a 6600 ohm load in push-pull, some at least 430V h.t. is required (taking an anode saturation dip of 80V) - this is irrespective of the valve used.  One is therefore limited by Ohms Law regarding maximum output power available.

So , basically , I should be looking at a valve that can handle higher anode voltage and/or current , so with valves I have available that leaves EL34 ( at a push ) , KT88 or GU50 . It seems to me that KT88 will be the easiest solution to achieve 30W , except for cost of replacement down the line . The GU50 is available at a quarter of the price of KT88 .

[Also keep in mind that valve data sheets give power available at the valve anode(s).  The efficiency of the OPT (at best some 90%) comes afterwards, thus a tabled output of 30W will give only 27W to the loudspeaker.]

Should I even be concerned by getting 30W output from the transformers , or would it be wise to settle for less power ? Will the transformer be happier ( perhaps last longer and have better bandwidth ) with less power ?

One more stupid question , but I prefer knowing to looking smart .  :giggle:

If the transformer is wound for 6600K : 4/8/16 , and most speakers these days are 4 or 8 ohm , would it be better to use the 8 and 16 ohm taps for 4 and 8 ohm ( 3300K : 2/4/8 ) ? So that one is using the entire winding for 8 ohm speakers , instead of half the winding ?

Thank you for your patience Ampdog - it is much appreciated .
“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Offline handsome

  • Trade Count: (+1)
  • AVForums Veteran
  • ***
  • Posts: 803
Re: Sansui 1000A Transformers on chassis
« Reply #9 on: February 16, 2017, 09:48:31 AM »
Transformer impedances might be fixed but loudspeaker impedances are not and waver from model to model as well as across the frequency range. Loadlines are therefore just approximations (in reality they are two dimensional ovals not one dimensional lines) ergo you needn't worry too much about being exact with your impedances. You can in theory mismatch your loads (e.g. 4R to 8R tap) to achieve different impedances but in a well designed transformer you will be compromising performance. Any class AB amplifier will be operating in class A at lower powers and yes the amount does depend on how much bias current you are using with the limit being how much power the valves will dissipate at idle. Therefore using lots of current will demand less HT and thus less power. In your case output transformer and the power transformer are fixed so despite using different tubes you are rather limited in output power. So don't scratch your head too much about that. Either triode or pentode operation (or make it switchable), fixed bias, 420V HT and design yourself some linear, robust and mean driver stages - have fun!

Offline Ampdog

  • Trade Count: (+1)
  • AVForums Veteran
  • *****
  • Posts: 6,695
Re: Sansui 1000A Transformers on chassis
« Reply #10 on: February 16, 2017, 03:29:07 PM »
Should I be be aiming at Class A PP ? Can one have the choice of Class A or AB1 in the same circuit depending on bias ?

As Handsome said.

The class AB working point is however chosen to give a smooth cross-over from class A to AB, not staying in class A 'long enough' to try give virtual class A performance at low output. (This is sometimes misunderstood but that is a design matter for another time.) But in the end not worth sweating over as Handsome said.

Quote
Should I even be concerned by getting 30W output from the transformers , or would it be wise to settle for less power ? Will the transformer be happier ( perhaps last longer and have better bandwidth ) with less power ?

Firstly some perspective: In reality a doubling in ouput power (or halving, the other way round) indicates a loudness difference of 3 dB, which is normally taken as just audible; not worth 'writing home' about. Thus it is academic to stress over output steps less than two times.

The bandwidth of OPTs is not (should not be!) dependant on the maximum allowed power rating.

With what you have I would settle for using the voltage doubling power supply arrangement, 6L6s and the existing OPT. Should give close to 30W output with the mains being full.

Quote
If the transformer is wound for 6600K : 4/8/16 , and most speakers these days are 4 or 8 ohm , would it be better to use the 8 and 16 ohm taps for 4 and 8 ohm ( 3300K : 2/4/8 ) ? So that one is using the entire winding for 8 ohm speakers , instead of half the winding ?

Yes .... BUT: In many OPTs the 4 ohm option is a tapped 8 ohm winding, often yieldieng less than optimal coupling to the primary, often also compromising NFB stability etc. I never consider that as an alternative unless secondary windings are brought out and can be arranged so that all windings will be used for any impedance. Doubling the primary impedance for triode operation by connecting 8 ohm loudspeakers across the 4 ohm outputs ... then the load impedance (13,2 K.ohm) is rather high again. Actual benefit becomes vague, also in the light of a varying loudspeaker load as Handsome outlined. One is dwelling in the mostly academic region.

That said, you can certainly also use the GU50s (weird looking devices - but since you have them ....)
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)

Offline Timber_MG

  • Trade Count: (+2)
  • AVForums Veteran
  • *****
  • Posts: 6,315
Re: Sansui 1000A Transformers on chassis
« Reply #11 on: February 16, 2017, 04:13:19 PM »
Subscribed, as I have a 1000A that was passed down to me.

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #12 on: February 16, 2017, 08:57:59 PM »
Thank you Ampdog and Handsome for your replies , and welcome aboard TimberMG .

I reckon I may as well take your advice Ampdog re the 6L6's and get on with it , and enjoy it as Handsome recommends . I'll see if I can come up with a schematic soon and we take it from there .

I love learning the theory though as I truly don't understand completely . Simultaneously , I must say that the practical experience of sorting the Quad 2's out taught me a lot with transparent results , which goes a lot further than theorizing - just seeing the proof in the pudding . :) 
“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Online fredeb

  • Trade Count: (+13)
  • AVForums Veteran
  • ***
  • Posts: 851
  • Lost in Music
Re: Sansui 1000A Transformers on chassis
« Reply #13 on: February 26, 2017, 04:43:22 PM »
Alright ,today I'm figuring out what the winding ratio of the output transformers are .

https://youtu.be/nNlPgirHP48 Uncle Doug - Output Transformers: A Discussion of Basic Principles

I'm attaching  33.7 VAC on the primary side ( Anode to Anode ) ,and getting 1.3 VAC and 1.85 VAC on the secondary taps . 33.7 VAC being the secondary tap provided for negative ( grid 1 ) supply .

So .....

33.7 : 1.3 = 26 : 1         winding ratio for tap closest to ground

Impedance = winding ratio squared x impedance

1 ohm : ( 26 x 26 ) x1  = 676
2 ohm : 1352
4 ohm : 2704
8 ohm : 5408             Alrighty , a seemingly useful ratio

33.7 : 1.85 = 18.2 : 1         winding ratio for tap furthest from ground

Impedance = winding ratio squared x impedance

1 ohm : 331.24
2 ohm : 662.48
4 ohm : 1324.96
8 ohm : 2649.92        Not to useful , methinks

Please let me know if I calculated incorrectly .

Here's a pic of voltages from power transformer , unfortunately I get 6.9V from the filament taps with 240V applied . I know from the schematic that 2 of the filament taps are meant to supply 2.5A and the 3rd 5A . I am assuming that the 2x yellow taps are the 5A taps , but i could be wrong .



« Last Edit: February 26, 2017, 05:00:06 PM by fredeb »
“And those who were seen dancing were thought to be insane by those who could not hear the music.”
― Friedrich Nietzsche

Offline Ampdog

  • Trade Count: (+1)
  • AVForums Veteran
  • *****
  • Posts: 6,695
Re: Sansui 1000A Transformers on chassis
« Reply #14 on: February 26, 2017, 10:57:47 PM »
....... unfortunately I get 6.9V from the filament taps with 240V applied .

YIKES!!

If you are certain that the primary winding is for the highest voltage, I fear a primary series resistor is indicated; 6,9V heater supply will shorten valve life. Yes, this is with no load, so better to determine later when all voltages are loaded, but I would guess that it would still be unacceptably high. (I begin to frown when 6,3V heater voltages go over 6,6V.) A series primary resistor is undesirable but unfortunately often the only way out.  A better way but cumbersome and unsightly, is to use a small bucking power transformer, bucking say 10V in  the primary circuit (small; only a few VA required, but less wasteful heat generated).
Audio must be the only branch of engineering where lack of basics' knowledge is considered a superior form of wisdom. (Anon)