• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

under-run push pull

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
After a bit of advice here guys. Heaven knows I'm no designer but I did have a fair amount practical knowledge under my belt a couple of decades back..

Basically I'm looking for a design with the same sort of ethos as the Quad 2, mostly from the viewpoint that they were bomb proof and under run the KT66's so they lasted forever:)

I'm after push pull, in the 12w/channel range and all tube, no semiconductors at all. Doesn't have to use KT66's of course but I would like nice big power tubes that are given an easy life.

Any ideas as so far I've come up blank. Most designs seem to like to push their valves close to their limits.. Id clone the 2's except for those dratted custom output transformers...

Thanks
 
You could add fixed bias if it has cathode bias and bias it as low as possible where it still sounds good.

Look at the EFB on on diytube for the Dynaco SCA/ST-35 for example. There is a link to it on the Tranola site. In a nutshell it allows fixed bias on any of the EL84/6V6/6973/7591 amps without getting another transformer. The voltage reg is not in the signal path.

Dave Gillespie the developer of it found you can lower the bias to about 27mA which gives 9W instead of the 12W design rating (for EL84/6BQ5) on the SCA/ST-35 amp for long tube life and the ability to run some new production tubes that normally wouldn't last.

Even though it was lowered from 32-35 mA on the stock amp the greater efficiency of fixed bias still allowed a increase of 3-4 W/CH with improved sonics to boot.

See the 5th post for the Tranola site with Dave's paper.

http://www.diytube.com/phpBB2/viewtopic.php?t=4374&start=0&sid=068430f593d40ae32382958fdc9ad683

A voltage tripler is another low cost way to add fixed bias if you have 5V on PS trans. Copied from Gordon W. at AK (audiokarma)

Hope that helps!

Randy
 

Attachments

  • voltatge_tripler 2.JPG
    voltatge_tripler 2.JPG
    11.8 KB · Views: 447
  • Voltage Trippler bias.doc
    23 KB · Views: 48
Last edited:
After a bit of advice here guys. Heaven knows I'm no designer but I did have a fair amount practical knowledge under my belt a couple of decades back..

Basically I'm looking for a design with the same sort of ethos as the Quad 2, mostly from the viewpoint that they were bomb proof and under run the KT66's so they lasted forever:)

I'm after push pull, in the 12w/channel range and all tube, no semiconductors at all. Doesn't have to use KT66's of course but I would like nice big power tubes that are given an easy life.

Any ideas as so far I've come up blank. Most designs seem to like to push their valves close to their limits.. Id clone the 2's except for those dratted custom output transformers...

Thanks

I would advice to build (almost) original Quad schematic with few changes:
1) Replace KT66 with 6P3S-E, which are very good and dirty cheap.
2) Since its not easy to find transformer with CFB, use 25W UL 6K6 transformer (and switch output stage from CFB to Ultra-Linear mode).
 
Remember as long as you lower the plate dissipation to a lower one where it still sounds good you can use any amp. Fixed bias makes it easier to do, and plate voltages can be left alone.

Non adjustable fixed bias can be found on some amps also, but they can be made adjustable.

Just look at the tube data sheets for the max. dissipation of the tubes your using.

Frank's electron Tube Data sheets

Randy
 
I had wondered what actual difference the output transformer made to the Quad, I sometimes had the feeling that Peter Walker was just being different for differences sake :) rather like the idea of doing a straight UL version...

Is it right then that the only thing that really makes a difference on valve life (given B+ being with tolerance etc) is bias?

A friend built a Mullard 5/20 but modern valves seem to last no time in it..
 
Yes.
The Mullard 5-20 has cathode bias, and tubes were cheap back in the day. Todays higher line voltages wouldn't help either. You could put bigger resistors to lower the bias on the 5-20 or other cathode bias amps, but it is a balancing act, so fixed bias (adjustable) is the much easier solution.

Cathode bias was cheaper and no adjustments were needed for the consumer, along with cheap tubes (replacing regularly was expected) so was common.
 
Last edited:
I started a PCB design for using 6LU8 tubes in PP configuration. The tubes are still low cost. The design is not all glass, it has a diode bridge to cap, but tube diode for slow stat to feed B+ to the PCB.
It can use the Edcor transformer set for Millets big red PCB. I'm still scrounging parts and the raw single sided PCB's arrived to transfer toner etch. It should be in the 20watt/channel ball-park, either UL or triode mode selectable. I use floral foam to mock up the design prior to etching the PCB's to verify component spacings.
There are many tubes and designs to choose from, happy hunting...
 

Attachments

  • TRY2.jpg
    TRY2.jpg
    778.1 KB · Views: 298
I've laid my hands on a copy of the original Mullard Circuits for Audio Amplifiers, while the 5/20 isn't exactly what I was after stock, at least I've got really clear instructions and there's a fair amount on the net about it.
I do know though that this design runs the EL34's at pretty much their max for the time and therefore way beyond what many modern versions can handle. Any ideas on 'cooling' it down and under running? Maybe decrease the bias and drop the B+? I'm very happy to drop, even halve the output..
Any pointers welcome, as I've said, I'm no designer (or mathematician!)
 
The simplest option would be to increase the EL34 cathode resistors. Alternatively, do what I did and adopt mixed bias: a slightly smaller cathode resistor and some fixed negative bias. This means you can run the valves a bit cooler. Mullard run them at nearly 70mA cathode current; I use about 55mA (from memory). I use RFT EL34s and they seem fine.

Mullard used such a high current for two reasons:
1. to get the highest power and so show off their new valve design,
2. to allow for some bias shift at high signal levels so that cathode bias can still be used.
Using partly or wholly fixed bias reduces/removes the bias shift caused by signal.
 
The simplest option would be to increase the EL34 cathode resistors. Alternatively, do what I did and adopt mixed bias: a slightly smaller cathode resistor and some fixed negative bias. This means you can run the valves a bit cooler. Mullard run them at nearly 70mA cathode current; I use about 55mA (from memory). I use RFT EL34s and they seem fine.

Mullard used such a high current for two reasons:
1. to get the highest power and so show off their new valve design,
2. to allow for some bias shift at high signal levels so that cathode bias can still be used.
Using partly or wholly fixed bias reduces/removes the bias shift caused by signal.

That's a great idea, did you build the amps 'as is' other then that? Were there any particular tips problems or pointers that come to mind? (I understand it was a while back :)
Appreciate the help, it gives me a kick in the right direction :)
 
I just simulated the UL-circuit given in Tungsol 5881 data sheet (at last page):

http://www.mif.pg.gda.pl/homepages/frank/sheets/127/5/5881.pdf

I modified the circuit by having separate (by passed ) cathode resistors ( 820 ohms).

When the supply voltage is dropped to 400 V the anode dissipation is just 14 W ( Ia = 38 mA).
With 6k6 output transformer the output power is 23 W with just 0,5 % THD.
The required drive voltage for output tubes is 25,5 Vrms.

If you also like to see suitable circuit diagram for voltage amplifying stage and phase splitter, I can show those.
 
vinyl dragon said:
That's a great idea, did you build the amps 'as is' other then that?
Almost. The phase splitter and output were near enough standard, apart from the mixed bias. The input stage I replaced with half an ECC81, so less gain and less noise. I also used a silicon rectifier.

5-20 OPTs may be expensive because UL OPTs need to be reasonable quality if you want to avoid oscillation. Having said that, I used Sowter OPTs and still had to add anode-g2 snubbers on the EL34s. Without them I was getting bursts of oscillation on signal peaks.
 
I used a soviet 6F12P as a voltage amplifier/cathodyne because I have used this particular circuit and I know that it works well.
However 1k resistor should be added as a grid stopper to the triode section of 6F12P.
Ofcourse Dynaco-like 6AN8 can be used instead.

This circuit has 13 dB global negative feedback. This can be adjusted by R18.
The output transformer is 6k6 to 8 ohms. Max. output is some 22 W and anode dissipation of the output tubes is some 13 W only.

An externally hosted image should be here but it was not working when we last tested it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.