HT voltage supply issue

hi!

i have done a somewhat simpel rectification/regulation circuit for the plates of my stereo tube amp for my passive eq.

secondary winding: 180VAC 100mA
rectifier bridge (4 x uf4004)
filtering: 2 x 100uF
MOSfet regulation: 240Vzener (filtered with 47uF) biased IRF740 (capable of a LOT of A and W)
outputvoltage expected: 235-240V
load: about 35mA per channel

i have only one side hooked up while trying to make final adjustments/tests of the tube circuits.
before i had one tube biased a little cooler and have been kind of fine with supply voltage.
but now i see that the dc get lowered from about 235V to around 220-230V and isnt very stiff as i think it was before. dont know if this is after rising bias on one tube from 24mA to 32mA or if something broke down. when doning sweeps i get an assymetric square wave on the output with a rounded rise and a small overshoot on the fall part. bandwidth is about half i had before...

would you think my mains transfomer is to light for me? even if it is stated to deliver 100mA at 180VAC.

i can see that the zener reference is getting messed up if the raw dc starts to fall but is my load of 35mA enough to cause this..?

if i pull the output tube the supply is around 250VDC.

ive checked some parts for faults but havent find anything yet...

also been thinking about adding a transistor and small resistor on the output for current limiting.

any suggestions about this?

With 180Vac, no-load unreg voltage is 252Vac; this leaves very little margin for regulation. You should either revise your expectations in terms of regulated voltage (about 230) or use a 200Vac xfmr.

You need to increase the current through the ZENER's as you have less than 1ma with no load. change the 22k to a lower value try 4.75k. This will supply 4-5 ma to the ZENER REF. You may also need to increase the 180 volts from the transformer.
Duke

I'm not sure the zener current is the main  concern - that suggested current is likely to cook a 5W zener from 1W dissipation unless the leads are nicely heatsinked.

The issue as indicated is the lack of headroom - so better to modify zener from 240V to say 220V and test again - with a meter to check for AC ripple from lack of regulation, or a cro on the output to check for waveform distortion from poor B+ reg.

Mains voltage variation can also catch you out in the future, so try to have 10% mains voltage tolerance if possible (you would likely need a variac to confirm that , so possibly not practical for you).

thanks alot for your answers!! now i know why im crying like a baby.

that resistor is actually a 2.2K and i even tried to decrease it just a little bit to see what would happen. no luck unfortunatly...

the transformer im using is an edcor. i see they have a 240VAC 200mA transformer that is just alitlle bit wider on one side but could still fit my back panel. more heater current also.

im gonna do some fast loadline caculation and see if i can pull it off with a zenerV of 220V instead. however, then the other channel comes in to play and doubles the load...

Suggest you be careful about zener dissipation.

The power in the ZENER now is 240v*0.001ma=240mw, 5ma then 1.2W or 600mw each .
Double the 35ma output load to 70ma and the ZENER power remains the same. The FET power changes with the load and not the zener power. The gate drive is not unchanged.
Duke

I'm just raising the concern of zener dissipation, as the part used would be a leaded 5W, and probably soldered directly in series, which if the close pair are required to dissipate 1.2W (or the actual level of twice that using a 2k2).  It's hard for such small parts to dissipate that level of heat unless the leads on both sides of each part have some heatsinking other than free air.

5v333, it seems by the bridge diode snubbers that you are concerned about rectification noise?

You may find that a better 'noise' outcome is to remove those diode snubbers, and direct your efforts to deploying a tuned snubber directly across the 180Vac winding, and assessing your wiring layout from the 180Vac winding out to C4-C9.

Imho, those snubber capacitors allow noise currents to flow through to C4-C9 to a far greater degree than just the off-state junction capacitance of the UF4007, and leakage inductance energy related to the 180Vac winding is far better managed by a local minimal-loop-area snubber, compared to giving that energy an easier path through the diode snubbers in to a larger circuit loop that may more easily include your audio circuitry.  Better management of the 180Vac leakage inductance energy may also noticeably reduce any coupling over to the heater winding (and then invading the audio circuitry through the valve heater-cathode interface).

Tube circuits should not be sensitive to 240V-220V changes.

If yours is, examine the circuit.

trobbins said:

5v333, it seems by the bridge diode snubbers that you are concerned about rectification noise?

You may find that a better 'noise' outcome is to remove those diode snubbers, and direct your efforts to deploying a tuned snubber directly across the 180Vac winding, and assessing your wiring layout from the 180Vac winding out to C4-C9.

Imho, those snubber capacitors allow noise currents to flow through to C4-C9 to a far greater degree than just the off-state junction capacitance of the UF4007, and leakage inductance energy related to the 180Vac winding is far better managed by a local minimal-loop-area snubber, compared to giving that energy an easier path through the diode snubbers in to a larger circuit loop that may more easily include your audio circuitry.  Better management of the 180Vac leakage inductance energy may also noticeably reduce any coupling over to the heater winding (and then invading the audio circuitry through the valve heater-cathode interface).

Click to expand...

i had a rf oscillation in my circuit. cant remember if it was more than one freq but i do remember i had a major around 6mhz.
it could be seen everywhere. audio paths and all the way down to the rectifiers.
i tried with conventional filtering in a lot of places with some success reducing noisy ripple but could not get my hands around the wild rf oscilations.

i read about diode snubbering on different forums and thought id try it out.
dont understand it in detail but when the diodes turn off it creates oscilations in conjuction with its capacitance, the secondary inductance and capacitance and the load.
there seems to be ways of calculate this but i never got that far before slamming some generic values in there. oscilations gone! this had alot of effects on the sound of the amp as well. more sweetness and openness and less stiffness.

my mains transformer  has flying leads coming out of it, the 180vac leads are probobly about 10-13cm long, i have the feeling that if a snubber network is to be connected across the secondary winding it should be as close to the winding as possible... so i have ruled it out...

did some new loadlines last night.
gonna take 2ma away of the output tubes
change the plate resistors of the prestage slightly
and change zeners in psu

something tells me it wont work when doubling the load with the other channel.