[SmittyHalibut]

That does look very nice.  I've been chatting with Lil'Knight in DM.  Turns out, he's out of the boards but has more on order.  They'll be delivered in a couple weeks he said.  I looked at the low-end build cost and decided it out of my price range right now.  (yes, I know, I'm in the wrong hobby if that's out of my range... 

)  I'll spend a bit of time putting together a board for the tube design you posted here and see what that would cost.

 

[chinsettawong]

If you really want to go with the tube amp, you can just hard wire the circuit together.  I didn't use a PCB either.
 
Oh, about KGSSHV, if you want to build a cheap version, you can really make it without the voltage regulator.  I'm sure it'll sound very nice too. However, I'm not sure if Dr. Gilmore and Spritzer will agree with me on that. 

 
Wachara C.

 

[Kuba4]

Second pair DIY electrostats by AmarokCZ
 

 

[chinsettawong]

The headphones look really nice. Congratulations!

Wachara C.

 

[spritzer]

Quote:

 
What modifications have you made to this circuit (if any), Wachara?  Are your phones still operating with the (220vRMS * 1.7vPtP/1vRMS * 2 (full wave)) = 748vDC BIAS?  I notice you don't have any DC blocking caps on the inputs, is that still the case?
 
Do you know if anyone has already made a kit for an ES direct drive amp like this one?  I'm not in it for the money, I just want the amp and it's cheaper to make circuit boards in quantity than one offs. 

 
No DC blocking caps needed on the input as the two stages are AC coupled.  A couple of changes I'd make for anything more than a dirt cheap build is to add a trim in series with the 5965 cathode resistor to null out DC offset and you can also add some low value trims on each of the front end cathodes before they are joined up to adjust the front end balance.  100r should be plenty for that role. 
 
I could also regulate the PSU which is well worth it.  Simple resistor-pass transistor-zener string is more than sufficient. 
 
Quote:

If you really want to go with the tube amp, you can just hard wire the circuit together.  I didn't use a PCB either.
 
Oh, about KGSSHV, if you want to build a cheap version, you can really make it without the voltage regulator.  I'm sure it'll sound very nice too. However, I'm not sure if Dr. Gilmore and Spritzer will agree with me on that. 

 
Wachara C.

 
I'd never build anything full size that isn't regulated (or some odd vintage tube design) but the main reason to regulate is to make sure the parts aren't operated past their limits.  Even Stax put a regulated PSU in the SRM-300 as the parts are used right on the limit. 

 

[FrankCooter]

Quote:

Hi,
 
I haven't made any changes to the circuit.  But the bias voltage isn't as high as you calculated, I think.  I calculate out to be around (220 x 1.414 x 2) 622vDC.  
 
There was a group buy a while back on KGSSHV PCB.  I think it is a very good and easy to build amp.  You might want to contact Lilknight to see if he still has some left.
 
Here are a few pictures of my KGSSHV that I just finished building recently.  I really like it. 
 
 
 

Great casework Wachara! Very clean and understated, yet elegant at the same time. A difficult combination to pull off. Really, really nice!

 

[SmittyHalibut]

I was considering trimmers on the plate resistors, but I confess I'm still learning tube circuit analysis. Trimmers on the cathodes would allow me to balance the bias, right? Would both be beneficial, or would trimming the plate resistor not accomplish much?

 

[FrankCooter]

Quote:

 
What modifications have you made to this circuit (if any), Wachara?  Are your phones still operating with the (220vRMS * 1.7vPtP/1vRMS * 2 (full wave)) = 748vDC BIAS?  I notice you don't have any DC blocking caps on the inputs, is that still the case?
 
Do you know if anyone has already made a kit for an ES direct drive amp like this one?  I'm not in it for the money, I just want the amp and it's cheaper to make circuit boards in quantity than one offs. 

If you dig around, on this and other forums, you'll find that a fair number of simple tube based electrostatic amp designs are out there. Most either originate or have been modernized by Spritzer or Dr. Gilmore. There's also the Egmont, the Exstata, the Stax SRX, and some transformer based designs. A good place to start is the Broskie articles.
 
Having heard Wachara's headphones, I have the highest respect for his abilities.  Having built a couple of similar amps, there are a few things I did differently that I thought improved the basic design.
 
First, incorporate all of Spritzer's suggestions about regulation and trimmers.  An amp like this really needs all the help it can get to maintain proper balance and off-set.
 
I would up the HV rails to at least 350V. 400V is even better.   I would change the plate resistors on the output tubes to IXYS10m90 based CCS.  Output tubes would be changed to an ECC99 to better handle the increased plate voltage and allow for increased plate current, which will improve slew rate.  I'd eliminate the cathode bias on the output tubes in favor of adjustable "fixed" bias.  A 12AX7 has a lot of gain, but to get it reasonably linear, you'll probably require some global feedback. You might want to look at a 12AT7 for the first stage.
 
If I was going to produce a board run for an amp of this type , I think I'd use the Stax SRX design as a starting point.

 

[SmittyHalibut]

Forgive my ignorance, I'm still learning how to read and design tube circuits.
 
Quote:

Originally Posted by spritzer /img/forum/go_quote.gif
 
A couple of changes I'd make for anything more than a dirt cheap build is to add a trim in series with the 5965 cathode resistor to null out DC offset and you can also add some low value trims on each of the front end cathodes before they are joined up to adjust the front end balance.  100r should be plenty for that role.

 
The trimmer on the 5965 is just a single trimmer in series with the 650 that's there, just to adjust the 650?  As opposed to a trimmer where each side goes to the cathodes and the wiper to the 650 to balance between the two cathodes?  That would create a different current path between the two cathodes and therefore allow a voltage to form across the cathodes.  Is that what you're suggesting?  Or, just making it so I can adjust the precise value of the 650?
 
Similarly for the 12AX7, you said "low value trims on each of the front end cathodes before they are joined up."  Could I do this with a single 100ohm trimmer split across the 38k as described above?  100 vs 153k is pretty negligible so I wouldn't expect the voltage formed across those two to be very significant.
 
Quote:

Having built a couple of similar amps, there are a few things I did differently that I thought improved the basic design.
 
I would up the HV rails to at least 350V. 400V is even better.   I would change the plate resistors on the output tubes to IXYS10m90 based CCS.  Output tubes would be changed to an ECC99 to better handle the increased plate voltage and allow for increased plate current, which will improve slew rate.  I'd eliminate the cathode bias on the output tubes in favor of adjustable "fixed" bias.  A 12AX7 has a lot of gain, but to get it reasonably linear, you'll probably require some global feedback. You might want to look at a 12AT7 for the first stage.  
If I was going to produce a board run for an amp of this type , I think I'd use the Stax SRX design as a starting point.

 
What are you attempting to do by increasing the HV rail?  Just give me a higher output voltage for the ES 'phones?   Or are you trying to stay closer to the middle of the gain curve to keep it more linear?  Or something else?
 
When you say "use the Stax SRX design as a starting point," Google returns a ridiculous number of hits for that.  But the one that seems most referenced elsewhere is this: http://www.bonavolta.ch/hobby/en/audio/hdgsta.htm   Is that what you're talking about?

 

[spritzer]

You would use a 500R resistor plus a 500R pot on the cathode and then adjust the output WRT ground.  This removes the DC offset so better driver control. 
 
Those low value trim pots are just for fine tuning the balance of the tubes so they can be much smaller than the actual circuit values.  See here:
 

 
The higher rail voltages mean more voltage swing but that also means the amp will be more linear under normal conditions.  More headroom and all that. 
 
That is indeed the SRX design, you can also look for ESX here on HF for the modified version that Kevin Gilmore and I did back in 2010. 

 

[chinsettawong]

Quote:

Great casework Wachara! Very clean and understated, yet elegant at the same time. A difficult combination to pull off. Really, really nice!

Thanks FrankCooter for your compliments.  I didn't do much on this chassis.  I simply utilized what I had on hand. 

@SmittyHalibut:  I can't help you much with amp designs.  Please listen to what Spritzer and FrankCooter have to say.  They really know what they're saying. 

  Good luck.
 
Wachara C.

 

[12517]

Quote:

Hi Gary,
 
If you can buy 0.5 mm PCB easily, please use it as the spacer.  It's much better than plastic.
 
I've got your email and will send the files to you later today.
 
Best regards,

Wachara C.

 

[chinsettawong]

Hi 12517,

Can you at least let us know a little about yourself? Do you have a CNC mill?

As you can see, I have built so many different versions. Which version would you like to build?

Wachara C.

 

[12517]

Quote:

Hi Khbaur330162
 
Thanks for your info.
 
One of these days when I have the time and mood, I'm going to make another pair of headphones with all the dimensions similar to HE90. 

  
 

 
Wachara C.

 
Quote:

I'm making a new pair of headphones.  This time I try to make the headphones with as less copper as possible - I etched out all the useless copper.  This helps reduce the capacitance on the headphones and makes them easier to be driven by an amp on high frequencies.  It's still a work in progress, but here are a few pictures:
 

 

 

 

 

 

 
 
 

 
 

 

 

[chinsettawong]

I just want to show you guys just how much weight you have to pull on the diaphragm.  When I tried to glue my diaphragm last week, this was what I did.
 

 
The Mylar I used was 3 microns.  I used a total of 8 bottles each with 650 ml.  That's about 650 grams on each bottle.  Still the stretch was barely enough.  If I were to do it again, I would probably increase the weight on each bottle to 800 -1,000 grams.
 
Wachara C.