GK-71 - economical build.

Paul Barker · Post by **Paul Barker** » Thu Jul 04, 2013 7:48 am

I don't know, just asking. Is the hf noise from the smps common mode? I assume it must be. Needs finding out what frequency, I think Andrew said to me at Owsten it was perhaps 50khz not shure if I got that right. Other reading on smps's I have since done the fundamental could be as low as 25khz which would require a monster choke. If it isn't common mode noise your cmc is worthless as you would of course know. The other issue is the two individual big muther 3 amp dc chokes (which you need four of). Are these in keeping with a cheap project?

On the one hand if we can crack the noise issue from smps power supplies cheaply we may elevate them into a higher respect class. but will it be any easier than building a conventional power supply both HT and filament?

I used the smps ht as an experiment and for cheapness. But I don't think I would be the one to make it sound like one of my better power supplies. Someone might be able to.

But for a cheap project which is perfectly acceptable at home for all but the critical listening smps in the HT and in the filaments is fine for me.

I just think the mountain to improve it is vast. But if you manage to fix it I am all ears.

JamesD · Post by **JamesD** » Thu Jul 04, 2013 8:35 am

Paul,

The noise from the SMPS is both common mode and differential and some of it is switching noise and some of it mains borne that couples across to the output so it is a wideband hash both ways around.

Fortunately I have most of the bits in my spares box - although only 2 dc chokes of the right size... I calculate I need 70mH at 3A - which doesn't sound too bad until you cost them up! I'll need to hunt a good price for them - although I can back off that spec given the OPT tx spec that I am using...

It will be really interesting to see how it would compare to ac and how it would compare to Andrews current reg! It won't equal andrews but to see how close or far it is would be instructive...

Edit. Damn! Just realised I got the calculation for inductance the wrong way around and its a 14H choke at 3A... Might need to rethink how I do this!!!
J

simon · Post by **simon** » Thu Jul 04, 2013 12:35 pm

The datasheet suggests the filaments could draw as much as 3.5A?

JamesD · Post by **JamesD** » Thu Jul 04, 2013 12:45 pm

yeah - I always over design by 50% for filaments - you have to GO LARGE for DHT and DHPs! Hence the 22V 6A SMPS too.

Thanks for pointing that out though - it could have been a problem.

J

pre65 · Post by **pre65** » Thu Jul 04, 2013 12:50 pm

JamesD wrote:yeah - I always over design by 50% for filaments - you have to GO LARGE for DHT and DHPs! Hence the 22V 6A SMPS too.

J

The filament PSU I use for 833a is rated at 30A max.

simon · Post by **simon** » Thu Jul 04, 2013 1:05 pm

I didn't think you'd missed it James

, but thought it might just be worth mentioning. Rod Coleman recommends three times the VA for cap input filament supply transformers which caught me out initially.

ed · Post by ed » Thu Jul 04, 2013 4:09 pm

simon wrote:I didn't think you'd missed it James , but thought it might just be worth mentioning. Rod Coleman recommends three times the VA for cap input filament supply transformers which caught me out initially.

yessirrreeee, caught me out as well.......if my last set of Txs were wound correctly then even 3 times is a bit risky.........the coleman regs are greedy to say the least when it comes to the high current ones...and then there's the cap spec to consider, current wise..

the advice about using multiple caps and the bit about appropriately high ripple current was crucial ime

...oh waily waily waily

andrew Ivimey · Post by **andrew Ivimey** » Thu Jul 04, 2013 4:41 pm

just a hint of encouragement!

IslandPink · Post by **IslandPink** » Thu Jul 04, 2013 6:19 pm

JamesD wrote:Edit. Damn! Just realised I got the calculation for inductance the wrong way around and its a 14H choke at 3A... Might need to rethink how I do this!!!
J

What's the requirement for this value ?

pre65 · Post by **pre65** » Thu Jul 04, 2013 6:24 pm

JamesD wrote:
Edit. Damn! Just realised I got the calculation for inductance the wrong way around and its a 14H choke at 3A... Might need to rethink how I do this!!!
J

14H sounds too big for a heater supply choke.

Paul Barker · Post by **Paul Barker** » Thu Jul 04, 2013 6:32 pm

andrew Ivimey wrote:just a hint of encouragement!

Paul Barker · Post by **Paul Barker** » Thu Jul 04, 2013 6:34 pm

Sounds quite a task heating them with the "ideal" method.

JamesD · Post by **JamesD** » Thu Jul 04, 2013 6:36 pm

It is much too big if it is just to suppress ripple on the heater lines.

unfortunately that was not my only intention - I want them to isolate the filament supply signal wise from the filament so that the signal current doesn't see the filament supply as alternative route to earth... So the criteria is that the choke impedance is higher at audio frequencies than the cathode resistor, and in typing that I have seen the way out of this problem - the cathode resistor is bypassed by the cathode capacitor, so its not a large choke impedance against 900R that I need from the chokes but a large impedance against 50mR or something like that

hurrah serendipity! Now we are at something like 15mH

I can get 15mH at several amps for less than £20...

Thank you all!

James

JamesD · Post by **JamesD** » Thu Jul 04, 2013 6:37 pm

Andrew,

Thank you!

J

Paul Barker · Post by **Paul Barker** » Thu Jul 04, 2013 7:01 pm

JamesD wrote: Now we are at something like 15mH I can get 15mH at several amps for less than £20...

Thank you all!

James

Now we are back on planet earth.

The choke you spec'd would be the size of a 3 ton truck.

But on a serious note, is it not worth looking instead at AC heating and ripple cancellation strategy? I recon that would be cheaper and lighter. It is the way I would go. IT seems to me the anecdotes of 3 X VA of filament transformer spec for the Rod modules added to all these chokes makes this a mamoth weight and cost of iron.

AC transformer only has to be the size it needs to be to produce that voltage at that current without making magneto constriction noises, so about 1.5 times industry standard (industry standard being noisy because it is an industry that doesn't mind a little noise from a transformer).

audio-talk

GK-71 - economical build.

#16

#17

#18

#19

#20

#21

#22

#23

#24

#25

#26

#27

#28

#29

#30