The 500w ruide transforners i have i about 3 nF between primary and secondary. That is the reason that i dont use floating power supply.
With a C-C transformer it is doable.
With a C-C transformer it is doable.
He also wrote this up in Jan Didden's Linear Audio.😉
I had a bit of discussion with Scott Wurcer about how to do a better op amp than the AD797. He said he had considered it but the only real improvement that came to mind was faster transistors from a newer process, so you could push up the Gain bandwidth product.
All the new developments seem to be to reduce power, size, cost not improve performance, and Scott said yeah, that was the reality. I wish he was still around.
Yeah, I still analyse motorbike motors and chassis, despite that I will never race, or even ride seriously any more. Just habit.
I do intend to build or buy one last set of amplifiers. But it's for home theatre so lots of channels, means lots of floated power supplies is less attractive.
Best wishes
David
3nF is easily OK to put on the output of a 'good' amp and should improve stability a little. If your design objects, it is a bad amp.
Yes. This is the case for the applications I still have practical & useful input on ... Condensor Microphones.
We don't really need more performance, For MM RIAA, NE 5534 is STILL probably in the top 3. It's only bettered by NJM2068 (sadly now Unobtainium) and a couple of $$$ supa dupas FETs like OPA627. Of course the 5534/2s should be NOS Mullard/Philips/Signetics samples hand carved from Unobtainium & solid BS by Southampton virgins. 😊
My little project is to see if da supa dupa FETs actually result in better audible performance 😲 by comparing SPICE world with some measurements and predictions based blind listening tests on RIAA preamps in Jurassic times.
"My little project is to see if da supa dupa FETs actually result in better audible performance 😲."
You can do better with discrete JFETs as a front end to a BJT OpAmp, as you presumably already know, at the cost of a few extra parts.
What you may have missed while you beach bummed is the discussion in the "Art of Electronics" 3rd edition and especially the "X chapters" supplement.
Intended for their scientific instrumentation but relevant I suspect. Sub nanovolt/rt Hz and low current noise, check it out.
Best wishes
David
You can do better with discrete JFETs as a front end to a BJT OpAmp, as you presumably already know, at the cost of a few extra parts.
What you may have missed while you beach bummed is the discussion in the "Art of Electronics" 3rd edition and especially the "X chapters" supplement.
Intended for their scientific instrumentation but relevant I suspect. Sub nanovolt/rt Hz and low current noise, check it out.
Best wishes
David
The reason for me not to use toroid transformers in floating power supply amplifiers lies in the fact that i have a few 100 kHz audio bandwith and a common ground system in my mains supply. The amplifier, TV and computer are in separate mains groups means that i have ground loops of up to 100 feet. The switching power supplys in tv and computer have about 5 nF mains to DC or there is a direct connection as in the Computer PSU. 5nF breaks the ground loop for 50Hz but when you have a few 100 kHz they conduct. And at that frequency the inductance in the main ground cables begin to play a role.
That makes a great risk of self oscillations in the higher freqencies.
That makes a great risk of self oscillations in the higher freqencies.
Oscillations and lots of induced noise. Amps like the QSC get away with it because they intentionally roll off above audio. If you are trying to make an amp flat a decade or more above audio, the coupling between windings is going to bother you.
If this really were that serious, any Circlotron would also suffer. I'll have to check my Yamaha A-S3000 integrated amplifier how the manufacturer solved it.
Best regards!
Best regards!
In a n Ei or C-C transformer it is easy to separate the windings so they can be made with just a few pF in between.
This one is called R-core.
This one is called R-core.
The important capacitance for a Vanderkooy or TransNova amp is between secondary and chassis. ie the mounting.
Stuff between primary & secondary will affect ALL stuff. The usual built in EMI suppressors on IEC sockets are quite good at reducing the effect of RF on the mains.
Just a reminder that ALL properly conducted DBLTs conducted this Millenium or the previous have shown for steep cut bandwidth limitation at around 20kHz ...
for the few true golden pinnae who can actually tell the difference on music & speech, (there are loadsa wannabe Golden Pinnae who THINK they can but show random results on DBLTs), they ALL preferred the bandwidth limitation.
The previous Millenium tests including my own, used some vinyl & analogue mastertapes as source. For these, I can pontificate at length about why this is so.
But the DBLTs conducted this Century used digital stuff which is already bandwidth limited so those arguments don't hold.
No one has investigated this experimental FACT seriously ... but if you want a Unique Selling Point that actually improves the sound of your amp, put a brickwall filter on its input. Maybe QSC got it right 😊
Last edited:
Been there and done that in da 70s. All that work went into the Don't Recycle bin when we became Beta Testers for Mullard TDA 1034, the original for NE5534
Been trying to find this without success. Could you post a link please
You can isolate the power transformer capacitance with a common mode choke, but it doesn't look like anyone has bothered to do that. Most amps are stable with a cap load 1000x as much. The mains line filter Y-cap should return any leakage current to ground. So, it's a non-issue.