Initial thermal slide is unavoidable due to the JFET has to stabilize. The 120k reference resistor and the trimmer play their role because they are the ones setting the JFET current and converting I to V. The ref resistor better be low ppm and the trimmer better be good or even substituted in a final assembly with a low ppm resistor of the trimmer's final measured value.
Hi regal!
What is wrong of my PCB???
Can you show me please???
Thank you![/QUOTE]
Nothing yet, haven't started building it just wanted to know which thread we were to go to for support.
So I got may damaged regulator working again.
I want to understand if the problem is the voltage, the inductive load or the startup.
To do this I plan to test an inductive load below 300 V first. My idea id to build a dummy load with an OPT and a 5K resistor in series and run them at 300 V.
At 300 V it should pass around 60 mA.
If this explodes again, I would conclude that the regulator cannot stand an inductive load.
If it does not explode I will raise the voltage to 480 V and see.
Next step, replace the resistor with a preheated tube.
I think that the explosion could also be caused by my unregulated PSU having 514 V when unloaded
Does all this make any sense ?
Thanks,
Davide
I want to understand if the problem is the voltage, the inductive load or the startup.
To do this I plan to test an inductive load below 300 V first. My idea id to build a dummy load with an OPT and a 5K resistor in series and run them at 300 V.
At 300 V it should pass around 60 mA.
If this explodes again, I would conclude that the regulator cannot stand an inductive load.
If it does not explode I will raise the voltage to 480 V and see.
Next step, replace the resistor with a preheated tube.
I think that the explosion could also be caused by my unregulated PSU having 514 V when unloaded
Does all this make any sense ?
Thanks,
Davide
You mean 380V not 480V yes? Your error amp BJT is -400V PNP.
Yes it makes sense. Shunt regs like this one have a certain on board termination compensation and a loop gain phase margin. Some types of loads will very well prove prohibitive.
Jumper that 47R that feeds the LEDS cap so it will start up instantly, maybe its helpful.
Your experiment step plan sounds careful. Good luck.
Yes it makes sense. Shunt regs like this one have a certain on board termination compensation and a loop gain phase margin. Some types of loads will very well prove prohibitive.
Jumper that 47R that feeds the LEDS cap so it will start up instantly, maybe its helpful.
Your experiment step plan sounds careful. Good luck.
Keep us posted, your experiment will provide valuable info.
Wondering if anyone can similuate an inductive load in SPICE?
I have done so on LTspice before and now. Not recommended. Only member Disco had run a couple of regs for SE amp in this thread, maybe it was 2A3 if I remember well, said it worked, but abandoned soon due to heat. In general I have written here that this reg is conceived for tube phono and tube line level. Now with the boards people got, I guess its unavoidable many applications are going to be tried out because its easy. But don't expect to marry with all. Regs in general expect a range of suitable conditions. A consideration is the CCS mosfet when doing heavy duty load experiments. Offload voltage of raw supply, violent load swings, are demanding. Something like FQP4P40 would be an interesting substitution test. Don't know the outcome, just an idea.
My application is a SE OPT headphone amp, 20mA's per channel. Basically a tube line level but with a SE OPT. Do you think this is worth a shot as the lower current should help feasibility or do you think I should do parafeed? Can you run LT Spice with 20 mA and a 5k ohm primary OPT? 210V in 160V reg.
We can not really predict the OPT, trust me. Instead...
I found some pics Disco sent me in 2009 about supplying some SE tube amp. EL12 driven by EL8, both in triode. He said it worked alright.
He had made the first double BJT Vref back then, more troublesome, plus own etched boards and had IRFP9240 for CCS. I think your board takes TO-247 CCS too.
Maybe yours worth a shot since its low consumption, you never know. Hey, you got boards, easy work. If it blows a couple of cheap leds and a semi, you could say you at least tried.
Attachments
We can not really predict the OPT, trust me. Instead...
I found some pics Disco sent me in 2009 about supplying some SE tube amp. EL12 driven by EL8, both in triode. He said it worked alright.
He had made the first double BJT Vref back then, more troublesome, plus own etched boards and had IRFP9240 for CCS. I think your board takes TO-247 CCS too.
Maybe yours worth a shot since its low consumption, you never know. Hey, you got boards, easy work. If it blows a couple of cheap leds and a semi, you could say you at least tried.
Yea it would sort of revolutionalize the headphone amp world so its worth a try. But I will diffenately use sockets for the silicone and LEDs.
So I did some tests yesterday.
I managed to have 380 V output at 70 mA on a OPT. Tubes were pre-heated. When I arrived at the final configuration the setup was Frankenstein style, so I only played some music on a single channel for few minutes, and did not make any measurement. I only verified that the regulator had some headroom to work (I could put the voltage up and down). The input is around 420 V. As Salas told me, I shortened the 47 R resistor in series with the 100 uF cap, on the leds..
Maybe tonight more data.
Davide
I managed to have 380 V output at 70 mA on a OPT. Tubes were pre-heated. When I arrived at the final configuration the setup was Frankenstein style, so I only played some music on a single channel for few minutes, and did not make any measurement. I only verified that the regulator had some headroom to work (I could put the voltage up and down). The input is around 420 V. As Salas told me, I shortened the 47 R resistor in series with the 100 uF cap, on the leds..
Maybe tonight more data.
Davide
that is ok!
Last I use SSHV for 6c45, V in 200V v- out 160v, Ia= 20mA. Transformer 5K/600R.
the SSHV is very very good.I love it
Right!
In my experiments reducing reg to load distance (for example from 20cm to 10cm) is really worth it. With such short cabling you can also try to remove any B+ direct bypass capacitor on amplifier side and see how it sounds.. but there is a possibility of instability.
Back to the thermal drift issue. I obtained a few 5ppm/degree 120k resistors. But it seems the best I can find for the trim pot is 50ppm, don't know if this will over-ride the benefit from the expensive 120k resistors. So I will replace the trimpot with a precision resistor after the build/setup. Curious as to what the wattage rating should be for the voltage adj trimpot/resistor ?