HT SMPS

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
ruffrecords said:
How do you know this? I it very hard to see the top side PCB tracking so I have not yet traced out the path the the opto isolator.
Magic ;). It's obvious how it is designed if you have the picture of the bottom PCB side.
 
ruffrecords said:
How do you know this? I it very hard to see the top side PCB tracking so I have not yet traced out the path the the opto isolator.
This is consistent with how PC power supplies work:  I once tried to use a small one for it's +12V and -12V rails, but it would not regulate properly without the 5V and 3.3V rail being loaded (even though I wasn't even using those rails).
 
Ive had a similar issue with a switcher , +12v -12v and 5 volts  ,unless the 5 volt rail was properly loaded with somewhere close to the designed current the 12 volt rails were low . with no load on the 5 volt line there was all kinds of rubbish and voltage was fluctuating all over the place . 
 
OK, so I will try a couple of EL84 tubes across the heater supply and see what happens. Presumably it would not be too hard to modify it so the HT rail is the one regulated?

Cheers

Ian
 
I just wired a couple of  Russian 6N6P tubes across the 6.3V winding which should draw about 1.5 amps. HT has now gone back up to 264V and the ripple is down to 20mV pp with a 35mA HT load. There is a (larger) 1/35mS component and a 1/28uS component so it looks like the switching frequency is about 35KHz.

There is clearly a compromise in multi-rail SMPS where only one rail is regulated and the others do the best they can with what they are given. Seems like a strong argument for a a single supply HY SMPS. I will try to trace the circuit  to to opto coupler to see if I can mod it to regulate from the HT supply instead.

Cheers

Ian
 
This kit arrived in the post this morning:

https://www.aliexpress.com/item/32969027567.html?spm=a2g0s.9042311.0.0.3da24c4dR5e85t

it is a basic push pull PWM SMPS based on the SG3525A chip. The blurb says the design operates at about 20KHz even though the chip runs up to 400KHz so I am tempted to tweak it to move the oscillator well out of the audio band. With this I should be able to build a reasonable starting point for an HT SMPS.

Cheers

Ian
 
Tubetec said:
Excellent plan , shifting  the smps dirt well beyond the audio spectrum .
I got the SG3525 board partly populated last night. Enough to get the oscillator going. Got it set at 111KHz.

Cheers

Ian
 
At different frequency, you may need to use a different ferrite core material to get any sort of efficiency.

Upping freq on the wrong type core usually leads to lots of wasted power that will go into core heating..

/Jakob E.
 
gyraf said:
At different frequency, you may need to use a different ferrite core material to get any sort of efficiency.

Upping freq on the wrong type core usually leads to lots of wasted power that will go into core heating..

/Jakob E.

You are absolutely right. The original design is supposedly for 500W output and it has a pretty big transformer. I do not expect to need more than 100W initially. That would gibe me 300V at over 300mA which is enough to power 25 of my tube modules which is just about right for the Mark 3 8 tracker I am building. So I am pretty sure I can go down a core size or two and at that point I will look more closely at core material.

For now I am just interested in getting something going.. This evening I added the FETs and their related components and soldered in the transformer. When I powered it up I got a nice chunky 55KHz square wave on the secondary. I just now connected a 1F4007 and a 100uF 440V cap across the secondary. As I powered it up I could see the current on my bench power upply juno up to about 3 amps then drop down to about 150mA as it charged up the cap. Mulitimeter showed 402V dc across the cap. Next I need to add a load and see what the ripple looks like.

Cheers

Ian
 
After some scary moment when I accidentally used a 1N4007 diode instead of a UF4007 (which resulted in several dead diodes) I have now got the thing basically working, I successfully connected the UF4007 to 100uF 400V cap plus 35mA load and ripple monitoring network. Result, no dead diodes, close to 225V dc and very short bursts of 50mV pp ripple every 8uS or so.

Output power is about 225 x 35 / 1000 = 7.875 watts

Input power is 12V at 0.75A = 9 watts.

So efficiency is 100 x 7.875/9 = 87.5%

Not a bad start.

Cheers

ian
 
Pretty good start, agreed.

Your short bursts of 50mV pp ripple every 8uS - what frequency would that be? I.e. if it's ultrasonic, could it be killed by simple L/C stage?

Jakob E.
 
gyraf said:
Pretty good start, agreed.

Your short bursts of 50mV pp ripple every 8uS - what frequency would that be? I.e. if it's ultrasonic, could it be killed by simple L/C stage?

Jakob E.

Yes, it is the switching frequency of about 110KHz. An following LC stage was exactly what I was thinking about next.

Cheers

Ian
 
Hey Ian, Jacob, and all other experts and experimenters:
I'm looking at using a SMPS with enough VA for a single channel's worth of valves, namely, 700mA of 6V3 heater, and around 20mA of 250V HT.
I was thinking of going the route of, say, a Meanwell, 8V5 4A low voltage SMPS, and then cobbling together a DC:DC boost converter for the HT.

I intend to use an appropriate LC filter after the 8V5 into a low dropout reg for quiet heaters and then, whatever is necessary for clean HT.
Since I'm a SMPS novice, am I off mark here or heading in a decent direction?
 
Meanwell are a good make so you are off to a good start in selecting them.

I always rate heater SMPS at twice the required heater current otherwise they often won't even power up into cold heaters and ones with hiccup mode current limit work better. On the other hand, do not over rate it too much else it may end up operating in hiccup mode because the load is very light and I know for certain this does stick crap on the heaters which definitely does find its way into the audio path.

Having said that, I now use Meanwell 12V SMPS for the heaters of all my mixers with no additional filtering. For 700mA your idea  of a linear reg is fine but you probably will not need the extra filter in between.

A dc/dc converter for HT is something I am looking into right now. I too have been thinking of generating it from the heater supply. However, winding this up to 250V or so is not so easy with a simple circuit. Someone has already done this in a 500 series module use a MAX something chip (can't rememeber the number off hand) - I will see if I can find the details. Anyway, I have designed a small PCB with this max chip to try it out. I expect the boards in the post any day now. However, I contacted MAX about this question and they put me in touch with an application engineer. Turns out the big problem is getting the ripple down to an acceptable level so it looks some additional filtering will be required there also. There are many, many other chips to choose from ( although I am limiting myself to DIP types) so I may just not yet have stumbled across the best one.

Cheers

Ian
 
Hey Ian,
thanks for the advice and insights, I'll take it all on board as I go along.
This is my first time using a SMPS with a DC:DC  flyback converter for H.T. voltages so I feel I'm in good company if you're on a similar path.

Until I build and test this I have no idea what (if any) L.T. and H.T. ripple I might have to tame or deal with in situ and my idea of an LC filter before a linear heater reg was really just a knee jerk decision.  A safety measure or 'get out of jail' card that might help stop any potential ripple from the H.T. feedback converter  from contaminating the heater supply.  Obviously, I'd prefer a lower parts cost so,  if I can do it with less stuff,  great :)

Have you tried any of the Meanwell supplies with lower VA than the big supply you're working on for your new desk?  I was looking at an open frame type to stick in the chassis and wondered if the bog standard ones rather than a medical type with 'Patient Protection' etc. is OK.  Proper safety measures in wiring up etc. being adhered to of course. 

Anyway, many thanks again for the advice Ian,
cheers,
David.




 
Back
Top