MJ Statistical Regulator PCBs
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#1 MJ Statistical Regulator PCBs
Seems to be some slight interest in these, so I'll make a batch up...
First draft PCB below - bit of a hack - early days - its minus the polyprop Kelvin-connected output cap but I'll do that tomorrow.
The jumpers allow you to monitor the screen & zener currents whilst adjusting VR1... Idea is that you populate the zener matrix starting at the top left and going down, then up, then down etc. There is space for 40 max. PCB is 5.5 x 4cm - If I can get it down a bit more I'd get 8 to a Eurocard...
TO-92 versions of the DN2540 are good for 120mA continuous...
When you have enough zeners for what you need, just place a jumper from the last one to the end of the matrix.
1 square inch (or inch square) of zener heaven Will have to look at the cooling so will probably space the matrix out a bit more... (about 4W going on in there...)
First draft PCB below - bit of a hack - early days - its minus the polyprop Kelvin-connected output cap but I'll do that tomorrow.
The jumpers allow you to monitor the screen & zener currents whilst adjusting VR1... Idea is that you populate the zener matrix starting at the top left and going down, then up, then down etc. There is space for 40 max. PCB is 5.5 x 4cm - If I can get it down a bit more I'd get 8 to a Eurocard...
TO-92 versions of the DN2540 are good for 120mA continuous...
When you have enough zeners for what you need, just place a jumper from the last one to the end of the matrix.
1 square inch (or inch square) of zener heaven Will have to look at the cooling so will probably space the matrix out a bit more... (about 4W going on in there...)
Vivitur ingenio, caetera mortis erunt
- Cressy Snr
- Amstrad Tower of Power
- Posts: 10582
- Joined: Wed May 30, 2007 12:25 am
- Location: South Yorks.
#2 Re: MJ Statistical Regulator PCBs
I'm certainly interested Nick
Sgt. Baker started talkin’ with a Bullhorn in his hand.
- Paul Barker
- Social Sevices have been notified
- Posts: 8991
- Joined: Mon May 21, 2007 9:42 pm
#3 Re: MJ Statistical Regulator PCBs
+1
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#4 Re: MJ Statistical Regulator PCBs
I've sourced some cheap 10uF 400V polypropylene caps that are reasonably small - they'll sit over the board quite neatly - should be about USD 5 each.
Any interest in doing a kit-of-parts for these - shouldn't cost very much at all... ?
Any interest in doing a kit-of-parts for these - shouldn't cost very much at all... ?
Vivitur ingenio, caetera mortis erunt
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#5 Re: MJ Statistical Regulator PCBs
Added the 10uF @ 400V polypropylene MKP cap - found one that is "only" 1.5" x 1", so not desperately bad.
Added a polygon for the output to get the best connection to the cap.
Wasted an hour shuffling bit around to see if I could make it smaller.
Added a polygon for the output to get the best connection to the cap.
Wasted an hour shuffling bit around to see if I could make it smaller.
Vivitur ingenio, caetera mortis erunt
- Paul Barker
- Social Sevices have been notified
- Posts: 8991
- Joined: Mon May 21, 2007 9:42 pm
#6 Re: MJ Statistical Regulator PCBs
Yes parts purchase good idea. I will need to arrive at 400v for the gk71. That polypropylene cap should do nicely.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#7 Re: MJ Statistical Regulator PCBs
With 40 x 5v6 zeners the board will only go to 224V - shouldn't use higher voltage zeners...
Thing to do would be to use two boards stacked - lower board is just for zeners - top board, leave out D1 and connect end of bottom board's zener chain to anode of top board D2. Just connect both boards'grounds together (pin 2) and it'll happily do 450V....
The design rules check is not setup for 400V working,so I'll use the UL rules and increase the isolation etc...
The cap is http://www.amazon.com/gp/product/B004H2 ... ge_o01_s00
Won't ship to the UK so had them sent to our NY office and they'll ship them to me...
Thing to do would be to use two boards stacked - lower board is just for zeners - top board, leave out D1 and connect end of bottom board's zener chain to anode of top board D2. Just connect both boards'grounds together (pin 2) and it'll happily do 450V....
The design rules check is not setup for 400V working,so I'll use the UL rules and increase the isolation etc...
The cap is http://www.amazon.com/gp/product/B004H2 ... ge_o01_s00
Won't ship to the UK so had them sent to our NY office and they'll ship them to me...
Vivitur ingenio, caetera mortis erunt
- Paul Barker
- Social Sevices have been notified
- Posts: 8991
- Joined: Mon May 21, 2007 9:42 pm
#8 Re: MJ Statistical Regulator PCBs
I see, thanks!
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- Cressy Snr
- Amstrad Tower of Power
- Posts: 10582
- Joined: Wed May 30, 2007 12:25 am
- Location: South Yorks.
#9 Re: MJ Statistical Regulator PCBs
Yes I'd definitely be interested.jack wrote:I've sourced some cheap 10uF 400V polypropylene caps that are reasonably small - they'll sit over the board quite neatly - should be about USD 5 each.
Any interest in doing a kit-of-parts for these - shouldn't cost very much at all... ?
Sgt. Baker started talkin’ with a Bullhorn in his hand.
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#10 Re: MJ Statistical Regulator PCBs
Almost there on this - small question for those interested...
Getting hold of the right N-channel depletion mode FETs with 400V+ Vds is tricky in in PTH packages.... all are SMT devices. Having said that, the IXTP02N50D looks like a nice device...
Doing the thermal budget, the TO-220 variants are available at 350V Vds, but not higher - the TO-92 package is simply too low power (1W) for this application, so what do people feel about possibly having the FETs as SMT devices (TO-243AA/SOT-89 @ 1.6W)? Note that these packages still have very low maximum power dissipation - the 350V TO-220 (15W) is still very much the preferred option.
Boards will be black solder mask, 40mm x 55mm (probably no bigger), M3 mounting holes, terminal strip connectivity, gold plated. I just like gold & black, but if people would prefer red or blue, let me know.
Getting hold of the right N-channel depletion mode FETs with 400V+ Vds is tricky in in PTH packages.... all are SMT devices. Having said that, the IXTP02N50D looks like a nice device...
Doing the thermal budget, the TO-220 variants are available at 350V Vds, but not higher - the TO-92 package is simply too low power (1W) for this application, so what do people feel about possibly having the FETs as SMT devices (TO-243AA/SOT-89 @ 1.6W)? Note that these packages still have very low maximum power dissipation - the 350V TO-220 (15W) is still very much the preferred option.
Boards will be black solder mask, 40mm x 55mm (probably no bigger), M3 mounting holes, terminal strip connectivity, gold plated. I just like gold & black, but if people would prefer red or blue, let me know.
Vivitur ingenio, caetera mortis erunt
#11 Re: MJ Statistical Regulator PCBs
Is the limitation for the fet the lack of a heat sink? The IXTP08N100D2 meets all I think you are after and its a lot less sensitive to static (IME) than the supertex parts.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#12 Re: MJ Statistical Regulator PCBs
That's a nice device, and cheaper (with better specs) that the other IXYS part I was looking at.
EDIT: I don't want to use heat sinks unless absolutely necessary, so the TO-220 is a good option - the heat sink is typically connected to the Drain, so is live...
EDIT: I don't want to use heat sinks unless absolutely necessary, so the TO-220 is a good option - the heat sink is typically connected to the Drain, so is live...
Vivitur ingenio, caetera mortis erunt
- Paul Barker
- Social Sevices have been notified
- Posts: 8991
- Joined: Mon May 21, 2007 9:42 pm
#13 Re: MJ Statistical Regulator PCBs
Looks good.
Much prefer soldering a 220 than sm.
Much prefer soldering a 220 than sm.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- Paul Barker
- Social Sevices have been notified
- Posts: 8991
- Joined: Mon May 21, 2007 9:42 pm
#14 Re: MJ Statistical Regulator PCBs
Jack I am not good at interpreting these SS device graphs. At what wattage do you think that we need to consider heat sinking the 220 package? Taking into account they are cascaded so splitting the dissipation between them.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- jack
- Thermionic Monk Status
- Posts: 5503
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#15 Re: MJ Statistical Regulator PCBs
I've got a bit of modelling to do in the thermal budget - Øjc (junction to case) is 8.3ºC/W and Øja (junction to ambient) is about 70ºC/W.
So, if you dissipating 1W, the case will be about 8ºC hotter than the junction and the junction will be about 70ºC hotter than ambient.
To do proper modelling is quite complex as ambient temperature round the case directly effects junction temperature, so free air movement & convection is important. Not only that, FETs are notorious for temperature sensitivity - specifically here is that RDSon changes significantly by 1.1%/ºC, so that has to be taken into account...
What is the maximum screen current you want to draw from this?
So, if you dissipating 1W, the case will be about 8ºC hotter than the junction and the junction will be about 70ºC hotter than ambient.
To do proper modelling is quite complex as ambient temperature round the case directly effects junction temperature, so free air movement & convection is important. Not only that, FETs are notorious for temperature sensitivity - specifically here is that RDSon changes significantly by 1.1%/ºC, so that has to be taken into account...
What is the maximum screen current you want to draw from this?
Vivitur ingenio, caetera mortis erunt