• 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.

DD amp for ESLs...

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
AX tech editor
Joined 2002
Paid Member
.... where are the pics ;)

This is what I was looking at, from the Fluke 5215. This is the equipment that has the low capacitance transformer.

Jan Didden
 

Attachments

  • output stage concept.jpg
    output stage concept.jpg
    38.7 KB · Views: 329
Hi Jan,

this cicuit looks very interesting!

If I guess right, It uses an optical coupling to span the voltage range... This could be a very interesting solution for solid state!

Wich led (sender) and diode (receiver) are used? Wouldn't this introduce an horrible amount of distotion and noise???

This approach could allow to cascade Fets so that a pass style amp for HV could be made....

Anyway, I think I otherwise don't really understand this circut... ;-)

Could you give some explanation??? (I'm so curious!!!) What is it used for??? You have some technical data???

Dirk
 
There now is a lpf from cathode to heater. So, with increasing frequency, I would expect to see an increasing voltage difference between the upper tube cathode and its heater. Depending on the allowed cathode-heater voltage, one must take care that the lpf (10k resistor and the xformer cap) does not cause this max to be exceeded at max frequency and max output level.

first I want to comment on Jan: I guess, that Dicks solution will exceed the heater cathode rating at HF, but luckily level will be low and nothing goes wrong...

OK, lets imagine a heaterxformer with 100pF capacitance between used windings and ground.
I connect the heater with the mentioned 10k to the cathode.
This means that the heater and the heaterxformer windings will follow the signal present at the cathode. At 20kHz the 100pF is equivalent to 80k ohm. This means 1/9 of the 20kHz signal level will be the cathode-heater voltage. At full output this is max 555V which is to high for heater-cathode.
But in music there will never be 20kHz signals present at full output, even above 10kHz.
Besides, the LPF at the input will decrease high frequenties substantial.
So my solution is safe to use.

Dirk, if I read your schematics right I see that when using the PL519 ( or 509), the signal should be applied to G1 ( which needs to be DC adjusted between 0 and –50V for correct bias ), and G2 is fixed at +50V.

Dick
 
AX tech editor
Joined 2002
Paid Member
djmiddelkoop said:




OK, lets imagine a heaterxformer with 100pF capacitance between used windings and ground.
I connect the heater with the mentioned 10k to the cathode.
This means that the heater and the heaterxformer windings will follow the signal present at the cathode. At 20kHz the 100pF is equivalent to 80k ohm. This means 1/9 of the 20kHz signal level will be the cathode-heater voltage. At full output this is max 555V which is to high for heater-cathode.
But in music there will never be 20kHz signals present at full output, even above 10kHz.
Besides, the LPF at the input will decrease high frequenties substantial.
So my solution is safe to use.

Dirk, if I read your schematics right I see that when using the PL519 ( or 509), the signal should be applied to G1 ( which needs to be DC adjusted between 0 and –50V for correct bias ), and G2 is fixed at +50V.

Dick


Yes, I can follow that. And you could also play with the resistor value depending on the transformer you use.

Thanks, I now have an alternative to all those expensive low cap transformers!

Jan Didden
 
Hi Dick,

I think you looked at the mu stage and the beta follower... Here I drive g2 (screen) and want to adjust idling current with g1. Looking a the synola 509 that seems to work... Is there a reason not to do so???

If you take a look a pics A, this should be your proposal??? Is this correct?

Regs, Dirk
 
Dick, you're right with schematic A. I'm stupid. I also have to keep the lower valve g2 on 50V. I think, the diff pair current source can be made adjustable to set the bias properly... Would this work?

Regs, Dirk

P.S.: @all: Which of the proposed circuits would you choose? And for what reson?

P.P.S.: Please don't mind that I forgot in B, C the connection of g3 with the cathode...
 
"this cicuit looks very interesting!"

The circuit Jan posted uses a push-pull (totem pole) emitter follower (Q11,Q12) to drive the output. The tubes provide tracking rail voltages for the collectors of the followers. (V1,V2 are a simple voltage boosted tracking cathode follower, and V3,V4 are plate following by means of the optical feedback and Op amp controlling the V3,V4 cathodes) Ie, the tubes just eat up the excessive voltage drops from the fixed supply rails (+/- 2.5 KV), leaving maybe 10 volts collector voltages for the Q11,Q12 followers centered around the output voltage (Vout +/- 10 V). The Q4 Jfet is likely used because of the limited pullup capabilty of the Q17 30 mA CCS load for the preceding (not shown) driver tube.

The high transconductance of the emitter followers provides a low output impedance and accurate drive (but is probably supplemented by a feedback loop unseen here).

Don
 
"Do you have an opinion on my HV amp circuit proposals?"

Well, I haven't used any of them, so I can't offer any expert opinion here. From MJ's book, the SRPP seems to offer the best pull-up and pull down capability, but also the highest distortion. The Mu follower is low distortion, but limited pull down, and the Beta follower is similar, but saves some voltage drop.

I would myself go for an updated version of Jan's schematic. (Its from a Fluke calibrator I think, PPM dist.) Instead of the HV tubes used to get the floating rails there, I would use switched CCS's top and bottom to charge a floating capacitor(s) for a floating low voltage amplifier. (Also, a small capacitor to ground is required to lower the impedance of the rails with respect to ground) Use two opto isolators similarly to Jan's scheme to control the voltage(s) there by feedback. Each switching CCS (top and bottom) would consist of two flyback xfmrs (optimised for the voltage/current requirements, and have to be able to provide peak drive currents) with a Mosfet on each one driving 180 degrees out of phase so as to get smooth, low ripple, current from the summed flybacks (HV rectifiers on output before summing). These would be PWM (pulse width modulated) to vary the current, so just use some standard PWM control chips to run them.

With the floating low voltage rails then available, you can put in most any type of amplifier design to actually control the output (totem pole tubes too even). The advantange here is that you can now use high gm parts at the lower rail voltages to control the panels with low output Z. (take a look at Broskie's latest balanced Aikido circuit for a tube drive idea: www.tubecad.com)

Obviously, a significant design effort though.

Don
 
Hmmm, here is a simpler way top get floating rails. Just use a conventional power amp to drive an ESL transformer. The HV output of this then drives floating +/- power supplies (needs low common mode capacitance to ground obviously, best way would again be a switching supply thru a well isolated ferrite xfmr. to get the low voltages) to power the output stage.

Don
 
Hi Don

you have a schematic of the first proposal?

But I fear, I don't understand. Finally, I understood the Fluke design. But: what is really strange with the Fluke for me, is the optocoupler: If I get distortion from the optocoupler, how can I get a good result finally??? Which devices are used for it (you may have a link with technical data?)

In this area I think of another solution:

Built a pass aleph like amp, but different in that way, that each output fet is optically coupled: The power for the receivers is retrieved from "light boxes" (solar cells) giving the energy...
Each fet is driven from the optical signal. An opamp does amplification to drive the gate...

Back to the original question: The beta-follower has limited pull down, because of it's constant current source? In this case the mu stage shouldn't suffer this problem - or???

What I'm not really sure about:

SRPP must be optimized for given load (symmetry!)
beta follower should not care (lower tube always sees constant current !???
I'm not really sure what the mu follower does.

If I go for sufficient high idle current, a beta follower should work best? (class a operation is a "MUST" )

But: SRPP would allow class ab - right???

Dirk

P.S.: Hope what I wrote is not completely rubbish...
 
AX tech editor
Joined 2002
Paid Member
The optocoupler in the Fluke design is to get the lower tube to keep the negative supply within spec. It really doesn't make a lot of difference if that supply varies by a volt or more. So any 'distortion' from that opto doesn't mess with the output (except via the PSRR of course). The Fluke has 10Hz-100kHz at less than 0.01% at 1200VRMS at 200mA RMS...

The opto is a pin diode, that is why you see the integrator at the bottom. At that time, they didn't have opto's with the bandwidth we have now, so now it would be easier, presumably.

Jan Didden
 
I think, the diff pair current source can be made adjustable to set the bias properly... Would this work?

Yes, i’ve made the CCS adjustable.

Don, I agree with your description of the Fluke circuit, thanks. Indeed, it’s a smart way to do this.
But it’s comprehensive, and surely only for the well experienced.

Very comprehensive are your other proposals, therefore I would not recommend them, sorry.

SRPP must be optimized for given load (symmetry!)
Correct, i will post the PL519 circuit, shortage of time prevents me from posting it now.
This circuit has been used without feedback and sounds good.
Other tubes I recommend are the tubes Sasha uses.

But: SRPP would allow class ab - right???

No, it is biased in class A.

Dick
 
Hi Jan,

now I get the point. (sometimes things take longer... ;-)) Because it's the collectors voltage, that is tracked.... !!!

So the Fluke circuit could be changed in that way, that I use a series of Fets that "track" the Voltage. All controlled with optocouplers. What a beautiful idea... I only knew designs before that controlled a Fet series with a resistor ladder. This was bad design to me!

@Dick: Hopefully I find time to draw the PL519 circuit this weekend, so save your time... ;-)

Dirk
 
AX tech editor
Joined 2002
Paid Member
Yes but note that this is a feedback loop. There is a complete adjustment procedure to make the loop stable with that varicap you see down in the circuit. This is a pin-diode and nowadays optocouplers are much better, higher speed but I would be carefull with putting a whole string of them in your amp. Do you use a simulator?

Jan Didden
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.