What is the same but upside down? I will try what you suggest and check the simulation. I will probably not implement it if simulation turns out worse. I did originally built a number of spare boards up when I made the amp as per pdf above so I can experiment practically but the latest circuit is very different from the original and maybe not practical. If there is even a remote improvement in performance I will report it here so now lets play a little. Haven't done this in many years. I was bored on Saturday and opened MicroCAP for the first time in five years.
H
HAYK
When an amp oscillate you wont forcibly notice it since the oscillation occur at very high frequency,
most of the time it can keep being listenable, just that it heats more than it should be for a given output power,
and in other cases, particularly with a bipolar OS, it can sound as if it s distorted, but with an oscillation that closely
follow the signal without reaching its amplitude you cant always spot the thing at first try.
Anyway since i have schematics that are close i did a try with yours, and as expected it oscillate on the first shot.
Notice that i get exactly the same frequency response, so either MicroCap is too forgiving or my simulator is too
intolerant.
And btw, C7/C24 and R61 are perfectly useless, there s no need of a TPC compensation for this branch of the driver,
to the contrary that s extremly counter productive, what should be done is to limit the frequency response of this side,
that s what Hitachi did in later schematics than their original AN.
Also the driver stage standing current is very low at 1.15mA, that s due to the IPS current mirrors that do not provide
the necessary voltage drop for the drivers s bases.
Attachments
Last edited:
I did the mods without any PCB changes. I am listening to it and I hear nothing different. The simulated result did make for THD to change from the bad 0.000002 something to 0.000001 something % THD. I have removed all caps on the board which extended the BW to 10MHz and no oscillation. Anyway, it did kill a somewhat boring day. My wife was not very happy with the mess I left so this is my next and final task, that of cleaning up this room. I have increased the bias to run at 6w class A but even that I don't hear. I guess I am getting too old to hear any significant changes anymore. So that was an interesting three days. I think my next task will be to give the Wadia a little clean-up and wipe the grit off the connectors etc. Then polish the scintillas frames and suck the dust off the membranes. I am not happy or unhappy, just content. But hey guys, it made me revisit MicroCAP after years which, is like getting hooked on a computer game and reminded me of the days I had a job to do.
Hitachi s laterals have a Vth that is 0.1 - 0.15V depending of the devices dispersion, so 0.65V is often enough
to bias energically the OS, moreover if the supply voltage is high as Ids increase with Vds at a given and fixed Vgs,
as pointed by Bernhard it should be 75mA for a 50V Vds.
It is a 1N4007 not an old red LED, an LED has a completely different symbol. Was just the right volt drop to bias outputs at about 6 watt each. Now running 24W class A. Nice and warm and the silent variable speed fans kick in and regulate the temp to 51 degree Centigrade.
I am okay with these component value changes and removing one current mirror and it made a small difference in simulated THD. What you guys are pointing out is fabulous for something that worked faultlessly every day for 20 years.
These numbers are all simulated in MicroCAP12 that is why I said that even if it was two orders of magnitude worse it is still pretty acceptable for me. I have been listening to it for several hours now and it does not sound any different than before the mods. I managed to kill the oscillation at 20 MHz which was visible amongst the noise on a scope.
As I said I had a bit of fun over the last few days and Wahab gave some interesting and useful comments which made me do some adjustments and I am pleased with the final simulated result, thanks bud. Even the Scintillas sound great after a good vacuum cleaning, they where bought in 1995.
A good night to you all.
Just to touch on something about MicroCap, I measured the exact voltages and currents that it presented so maybe it is good for THD and frequency response as well.
I am okay with these component value changes and removing one current mirror and it made a small difference in simulated THD. What you guys are pointing out is fabulous for something that worked faultlessly every day for 20 years.
These numbers are all simulated in MicroCAP12 that is why I said that even if it was two orders of magnitude worse it is still pretty acceptable for me. I have been listening to it for several hours now and it does not sound any different than before the mods. I managed to kill the oscillation at 20 MHz which was visible amongst the noise on a scope.
As I said I had a bit of fun over the last few days and Wahab gave some interesting and useful comments which made me do some adjustments and I am pleased with the final simulated result, thanks bud. Even the Scintillas sound great after a good vacuum cleaning, they where bought in 1995.
A good night to you all.
Just to touch on something about MicroCap, I measured the exact voltages and currents that it presented so maybe it is good for THD and frequency response as well.
