audio-talk

Looks like the HV version isn't quite as high as I recalled, 200VAC input. Anyway, here are some links;

https://www.diyaudio.com/forums/group-b ... er-gb.html

https://evotronix.eu/

Mike did something similar (Without rectifier) some years ago.

Nick wrote: ↑Fri Aug 16, 2019 10:45 pm Mike did something similar (Without rectifier) some years ago.

Are you thinking of his 'E-Choke'?

Yep.

Nick's mention of Mike's E-choke project reminded me that I have a couple of the boards in my spares box - I might just dust them off and use them in my 300B HPA project. Must try to find the thread for them.

Nick's mention of Mike's E-choke project reminded me that I have a couple of the boards in my spares box - I might just dust them off and use them in my 300B HPA project.

It will improve things as it has been shown on the Tentlabs page but it's not the same as the SLB. I think the active bridge rectifier adds a big part of the wonderful improvement compared to regular rectifiers (even to Schottky diodes which I have used originally) with typical CRC/CLC.


I am still struggling a bit with finding the best solution for protecting the speakers from tube discharging or other faults caused by DC.
I studied at least 10 different circuits (among them the solutions from Neurochrome and XRK which are really the best compared to the rest) but nothing really suits the special situation of the InvOTL. A TVS diode would have been a perfect solution for the dischargings but is no longer an option because it ruins the transients/ sound in general.

The InvOTL needs protection from +/- 40V and -250V. The trigger level for the latter needs to be as low as possible (without being triggered by the DC offset!) in order to give enough time for the SSRs to cut it off before it can reach high enough levels at the output that could do damage to the speakers and/or the SSRs (60V upper limit). The rising time for the 40V is not so critical.
The cause for any DC problems at the output can only be a physical short in the tube (grid to anode for example) or a gassy tube.

I bought a new tube tester (http://www.space-tech-lab.com/Testing_E ... ter-2.phps) that can test tubes for noise and gassy/grid leak in order to get final clarity and indeed the one tube that had caused this discharging had this problem.Nice to see how the blue LED ("gassy tube") lights up and gives clarity. The crackling sound which was driving me crazy came from one triode of a new pair of Russian 6SN7 which showed no signs of any problem in the TS tube analyzer (original kit). First I thought it could also be a problem caused by another gassy 6AS7 or maybe a cold solder joint. The cause was an extremely microphonic triode, easy to find with the help of the audio output of this tube tester.

So testing all the tubes that I use in the InvOTL with the Space Tech tube tester should already be enough but I want perfect safety and not everybody who might build this amp wants to spend money for a new tube tester. Finding a potentially faulty 6AS7 in the actual amp without tube tester is quite an endeavor.

I boiled it down to two circuits one of them comes from Kelvin who passed away recently.
Here the question is how to set the the correct input sensitivity for +/-40V and -250V for two different scenarios? Is it possible at all? The 40V would create high current, the -250V very little current but high voltage. How would I get high enough levels for both voltages with one set of resistors so that they would trigger a clear logic state (0 or 1) without flickering?

The second possible circuit I will post later, it's not ready yet.

Any comments or suggestions regarding how to calculate the correct values for R1-R4 or change the circuit so that it would fit both needs?

can you run a spice sim with a 74 series logic library?

I don't understand, if you have a trigger at +- 40v, then that will also trigger at +250v? If the problem is the voltage limit of the solid state relay, why not use a mechanical one, and or a relay in the supply lines to remove the source of the voltage. Or a crowbar circuit maybe?

Wolfgang wrote: ↑Thu Aug 29, 2019 4:07 pm

Nick's mention of Mike's E-choke project reminded me that I have a couple of the boards in my spares box - I might just dust them off and use them in my 300B HPA project.

It will improve things as it has been shown on the Tentlabs page but it's not the same as the SLB...

I didn't think it was Wolfgang, I'm considering a different use case.

can you run a spice sim with a 74 series logic library?

I did only simple things with spice. Would probably need some time to get useable results.

if you have a trigger at +- 40v, then that will also trigger at +250v? If the problem is the voltage limit of the solid state relay, why not use a mechanical one, and or a relay in the supply lines to remove the source of the voltage. Or a crowbar circuit maybe?

With higher voltages it's all about how fast the relay can react in order to avoid damaging the speakers. Switching speed of the SSR between 0,7 and 2,5ms is good enough and only a little better than mechanical relays which would also work. The earlier it triggers for the -250V the less time for the voltage to rise to critical levels. So ideally it should be a very fast acting active circuit with no influence on the audio signal. In case of a faulty tube it should make no or only some hardly noticeable sound in the speakers and also wouldn't damage the SSR.
It’s not clear to me how to select correct values for R1-R4 so that it would work for +/-40V and -250V in order to set threshold levels for the circuit for fast enough reaction time and at the same time with the correct voltage for the input of the CD4093. Probably it's not possible.

Wolfgang, a question about your output stage power supplies. You currently use the SLB modules; these give you a very clean supply but are not regulators so I assume that you keep them stable, with respect to the output voltage, because you use a mains conditioner so your amp always has the same supply voltage? If so, in the absence of a mains conditioner, would regulated supplies be a good idea - I'm thinking of maybe the latest Salas BIB or Jung-Didden super reg, both of which can be configured for 40V I believe (so 11V shouldn't be a problem) at upto 1A.

Ray

and at the same time with the correct voltage for the input of the CD4093.

I assumed the diodes on the input would protect the CMOS from over voltage. And at worst, socket the IC and buy a bag of them and include a test button to bleed voltage to the output.

Wolfgang wrote: ↑Thu Aug 29, 2019 4:07 pm I am still struggling a bit with finding the best solution for protecting the speakers from tube discharging or other faults caused by DC.
I studied at least 10 different circuits (among them the solutions from Neurochrome and XRK which are really the best compared to the rest) but nothing really suits the special situation of the InvOTL...

Is the issue with the Neurochrome/XRK solution suitability down to them being unable to handle the 250V scenario or is there something else? I know the XRK board specifies that it can handle upto 48V of DC and assume the neurochrome will be similar as the designs are quite alike.

I assumed the diodes on the input would protect the CMOS from over voltage. And at worst, socket the IC and buy a bag of them and include a test button to bleed voltage to the output.

With this circuit it's clear: The DC level will be controlled by the diode bridge to impose a forward bias polarity to the B-E of the transistor. With Kelvin's circuit I cannot see how a DC condition would make the circuit work as there are no caps that would charge up and precisely define a DC level and I wouldn't know how to adjust R to levels that would work with 2 voltages and be highly sensitive to both voltages as there are more conditions to be fulfilled: 1.the applied voltage should never exceed the supply voltage to the IC (I need it at 12V)
2.Undefined voltage levels according to CMOS gates are within 0.75 and 2.5 volts.
Maybe I simply cannot see the obvious but as the diodes are referenced on one side to the supply voltage how could they control the input voltage/DC level as in the example above? Maybe this circuit doesn't work properly for a situation like the InvOTL .

I will use my other option which I have already tested. It works stable down to 1,6Hz and triggers for the neg side already at 0,8V.

You currently use the SLB modules; these give you a very clean supply but are not regulators so I assume that you keep them stable, with respect to the output voltage, because you use a mains conditioner so your amp always has the same supply voltage? If so, in the absence of a mains conditioner, would regulated supplies be a good idea - I'm thinking of maybe the latest Salas BIB or Jung-Didden super reg, both of which can be configured for 40V I believe (so 11V shouldn't be a problem) at upto 1A.

Yes, I use a PS audio power plant which makes it possible to use AC for the heaters and the SLB in case of more than 1-2V mains fluctuations.
Otherwise I would adjust the AC heater resistors for the highest possible mains fluctuation level and it's all ok.

I don't know what's more important or sounds "better". Originally I wanted to use the Salas BIB (got 2 PCBs) but their max input voltage is 40V and the 30V toroids have 43V after rectification. If you use tailor made transformers it should work. But you will need big heat sinks for this !
The SLBs on the other hand have an incredible influence on the final sound and I wouldn't want to give up even 1% of that. It wouldn't matter if the mains voltage fluctuates a little. You simply won't get the max swing of 16Vp-p into 15R. But you have so much power for every dynamic situation that it doesn't matter. My output stage operates now at 39V instead of originally 42V and there is no problem.