Precise passively smoothed bench HT supply (with total recall)

[Tubetec]

Its awkward to change series resistors and zeners to vary the Ht voltage off a regular passive mains transformer/bridge/filter supply for test purposes.

Instead , how about using your programable bench top arb gen to drive a standard step down transformer in reverse
and generate a variable HT voltage . Smoothing can be done in the traditional way , RC , but for the sake of 13 euros worth of a hammond choke(150H/8mA) , worth putting some Henries into the equation to make sure your reference voltage is down to a few uV noise .
The output of the gen can be trimmed under load and stored as a preset , you can dial in the anode voltage to a very fine degree of precision on a given tube and snapshot it if you want
Of course your not limited to the usual 50/60/100/120hz residual hum from mains opperation either , swap out the iron core for ferrite and square wave/saturate into the Mhz region if thats what tingles your fancy .

 

[Tubetec]

So I had a chance to bread board my idea ,
I way overestimated the drive capabillity of my ARB , First I went with a 2x20v CT :230v 20va transformer ,applying the drive with both channels of the arb , I was still only able to make 6Vrms drive under load so output voltages were low ,
By connecting half a 20v winding I got around 85volts DC ,
I tried again with a different transformer , this time 6v to 230 , I got around 95v at the filter output
Frequency of drive Ive tried varying from 10hz to around 1khz , it doesnt appear to make much difference appart from small changes in the DC voltage at the output . Sine wave drive gave much lower DC voltage than applying a square wave , but had no effect on the noise level .

I happened to have a nice mil grade metal encapsulated inductor , 7Hy , I made a choke input filter followed by 4 RC stages with 470uf 200v electrolytics ,Series resistors are 680 off the rectifier before the choke, then 10,12,4.7k , loaded with a 75kohms it draws around 1ma .It takes around 20 minutes to ramp up but once its there it doesnt vary more that a few tenths of a volt

when I make contact with the ground of the measurement system the noise falls close -92dbv, the calculated noise of the 75k load resistor comes out at -106dbv and I used low noise resistors at the end of the RC chain just in case . Ive often found high voltage electrolytics to contribute frying noises of one kind or another ,especially if not used in a while ,I was considering extra foil bypass caps on the electrolytics and a 4uf oil block capacitor at the end , or maybe a second choke , either way Im gone beyond the usefulness of the Fluke for noise measurement I'll try REW with a mini audio interface with a high voltage blocking cap and bleed resistor into Hi-z and apply a known amount of gain , then look and listen to the noise floor .

I'll have another dig around in the junk box and try and find a better matched transformer. Im already using half a 12volt winding and only getting 95 volts , 6volts is the lowest you find most off the shelf transformers ,where dual 3volt windings in parrallel would suit me better .

A low profile mains transformer with dual pri/sec windings is what I want but I cant seem to find them below 6 volts AC ,suggestions welcome .

 

[Tubetec]

With no load my Arb gen puts out 15,6volts RMS , I apply that voltage to a 15v winding of a tiny 1.2va transformer Initially the output of the arb drops due to load but as the caps on the HT side charge the load is gradually reduced I still only manage around 120vac off the 240 volt winding under load , that winding measures 4.2k ohms , am I loosing half my AC voltage across the transformer winding itself ? Anytime I drove two transformers back to back before I got more or less the expected voltage minus some small loss . With all the caps dissconected I manage about 145 v ac at the transformer output ,rect input . As I said previously when inputing a square wave I get around 85 volts , using a sine around only 60 volts dc comes out .
This transformer has two 15volt windings , currently Im only using one winding because Im a bit unsure about the pinout , the two 15v windings come out as 2 pairs of four pins , apparently marked 6-7 9-8 , obviously the actual pins on the transformer appear in two rows 12345-6789 , when I tried it previously I wired 6 to 8 and 7 to 9 which appears to put one winding out of phase . I'll be back .

 

[abbey road d enfer]

, how about using your programable bench top arb gen to drive a standard step down transformer in reverse

"arb gen" ???

 

[moamps]

"arb gen" ???

Maybe
https://en.wikipedia.org/wiki/Arbitrary_waveform_generator

 

[abbey road d enfer]

Maybe
https://en.wikipedia.org/wiki/Arbitrary_waveform_generator

Probably. Thanks.

 

[mad.ax]

am I loosing half my AC voltage across the transformer winding itself ?

A 2 X 15 > 240V transformer is = to a 30V > 240V one. If you feed it only 15 instead of 30, sounds right to me that you only get half of the 240....

Now if your 'arb gen' is not powerful enough, you could use some kind of audio power amp in order to beef the output up.
Obviously you'll need to monitor the pa output cause it will have some voltage gain

 

[Tubetec]

Im using the Feeltech Fy6600 ,
I swapped out the power supply for linear so it has a bit more drive than the stock version .

At the moment Im running the generator into half a centre tapped winding and getting around 150v at the rectifier output ,112v after the filter stages , the generator is maxed out .
Im using a 75kohm load but thats no so appropriate as most high gain tube stages use at least 100kohms load ,
I'll mock up a schematic on PSUD for clarity .

 

[Tubetec]

Ive simulated things a bit better here , voltages are correct ,but the actual value of resistances is 680 ohms , 12k ,10k, 4.7k and 100k load although I couldnt get the numbers to match up with those values in the sim ,

 

[JohnRoberts]

Back when I was designing my outlet tester, I need 500VAC for insulation testing. I had a bunch of 120x20VAC power transformers sitting around. I wired up the first one 120V input 20V output. The I wired up 4 more with all 5 of the 20V windings in parallel. The 4 floating 120V primaries got wired in series to make my near 500VAC output source... good enough for crude bench measurements. I put a 1/4A fuse between the 20V windings but it could still sting you good (bad?)

JR

 

[moamps]

Im using the Feeltech Fy6600 ,
I swapped out the power supply for linear so it has a bit more drive than the stock version .

The next step forward could be to introduce a control (NFB) loop into the VCO input of the arbitrary generator from the output.

 

[Tubetec]

Its probably possible to vary the duty cycle of the square wave via the VCO input but the output voltage is already very stable , Its tended to creep up a little over the last three days , but my guess is thats due to the caps not having been used in a few years and needing to re-form .

If I want to listen to the noise on the supply directly can I cap couple from the output with a high value resistor after to ground , so I dont end up applying volts DC to my audio interface ? Im guessing hi-z input might have less of a loading effect on any noise present but backround noise in hi-z inputs often tends to be more than the -92dbv noise level I read with the Fluke.

 

[Tubetec]

I found a good article here on choke supplies ,
http://www.emissionlabs.com/Article...igning-with-5U4G-and-274B-POWER-SUPPLIES.html