goldenreference low voltage power supply

[Pars]

Oh all right...

with std. rectifiers

 

With Schottky bridges

One other thing I noticed while playing around with this. Since this is just on a piece of wood, I don't have the AC ground tab connected to anything.

When checking AC voltage between it and the board ground, it was floating around at 50-60Vac. I assume this is normal since the board isn't really referenced to anything? Should I tie the AC ground tab to the board, either directly or thru the 10 ohm resistor that is provided?

Edited September 2, 2017 by Pars

[Pars]

Was playing around with input voltage and looking at dropout. This doesn't drop out of regulation untli input is ~16.8Vac or so (output is still at 19.95Vdc). DC at the filter caps is ~22.7Vdc at that time (IIRC). I thought it would drop out before that. This is completely unloaded. I think I will go with the 20V transformer, as Michael suggested. Even at 24Vac input, pass devices were not warm, though I'm sure that will change when it is loaded.

 

EDIT: trying to recall, do most builders tie the AC ground to the PSU or amp board grounds, or just to the chassis? Don't any RCA connectors, etc. in effect tie the board ground to AC ground/chassis? I know that XLRs wouldn't.

Edited December 7, 2016 by Pars

[spritzer]

I tie everything together and even in sensitive circuits like the Carbon this is a non issue. 

[mwl168]

On my builds, which are all balanced inputs with Neutrik XLR connectors, I tie PSU ground to the chassis ground (star ground) which is then tied to the AC ground. No ground connection between the input XLR and the amp's input. The input XLR ground pin is tied to the chassis as is suggested by the white paper "pin 1 revisited" by Rane.

For my Carbon which has the GR LV and GR HV PSUs, I run a heavier gauge wire between the ground of the two PSU and run one of the PSU's ground to the star ground.

Seems to work very well for me.

Edited December 7, 2016 by mwl168

[kevin gilmore]

pars needs an 8.5 digit dvm!

 

[spritzer]

We all need that!!!   

My own version of the GRLV is slowly being assembled on my bench and when they come out ok I'll convert the design over to SMD for some future projects.  Now I just need a pick and place machine...and solder paste applier...and an oven...   

[Pars]

2 hours ago, kevin gilmore said:

pars needs an 8.5 digit dvm!

 

I'm sure you have a few laying around Kevin... I don't live that far away

1 minute ago, spritzer said:

We all need that!!!   

My own version of the GRLV is slowly being assembled on my bench and when they come out ok I'll convert the design over to SMD for some future projects.  Now I just need a pick and place machine...and solder paste applier...and an oven...   

Surely you have a few ovens around Mr. Baker-man. Any old toaster ovens? Timer/programmable?

[wink]

12 hours ago, kevin gilmore said:

pars needs an 8.5 digit dvm!

 

He's close in model number with the HP3468A.

Needs to drop the third digit for the HP/Agilent 3458A/B

http://www.testequipmenthq.com/datasheets/Agilent-3458A-Datasheet.pdf

We used these to calibrate Fluke 5500, 5700, 5720 calibrators on the rsisitance and DC volts and current.

For the AC volts and current something a little more accurate like the Fluke 5790B or Wavetek/Datron 4920 were used.

[Pars]

Could I request that a second set of output terminals be added to the next revision? This would make things easier when feeding more than one board.


Sent from my iPhone using Tapatalk

[kevin gilmore]

not without making the board a bit bigger, its just about as small as I could get it

[Pars]

Could you:

1. move the power on LED and the 2K resistors
2. leave the mounting holes in the same places, but extend the board out a bit to allow moving the output connector(s) out a touch, allowing for a couple of them
3. also, is anyone actually using the adjustment pots? with good gain resistor measurement, you can get the output voltages quite tight without them (my board is a previous version without them). This would free up considerable real estate.

Just some thoughts. 2x18ga wires into a Phoenix block isn't much fun, and I don't think the Molex SPOX connectors I prefer using would accept more than 1 wire per terminal either, though I haven't tried it yet.

Edited December 23, 2016 by Pars

[kevin gilmore]

goldenreference6d posted, 2 x output power connectors.

one should be the short variety, and the other one the taller one, or the angled one.

board is now 3.95 x 4.32 inches

its better to not use the pots and resistors, it probably increases the tempco. if you hand select the feedback resistors, or buy .1% parts its a better idea.

 

Edited December 23, 2016 by kevin gilmore

[Pars]

Very cool, thank Kevin!

[mtoc]

Folks if I wanna adjust the output voltage, which resistor should I change?

edit: +/- 12V/13.5V/15V, all less than +/- 20V, should I also need to change the caps and ops?

Edited December 26, 2016 by mtoc

[ang728]

http://imgur.com/xKQhzZZ

Take from Pars's pic,Change those 1.5K resistor,top and bottom ones are R8,R9

for V+: (R8 + R7) / R7 x 10 (reference voltage of D5 - LT1021-10)

for V-: (R9 + R10) / R10 x 10 (reference voltage of D7 - LT1021-10)

For 15V you can go 750R,1.5K Set

others you may need to change different voltage version of  LT1021

don't need to change caps and OPAs for voltages you mentioned ,but you may need to change zenzer value to voltage/2

Edited December 27, 2016 by ang728

[mtoc]

edit 2: the transformer I think I will choose 12vac, 15vac or 18vac (all *2 models), among these three.

[ang728]

49 minutes ago, mtoc said:

edit 2: the transformer I think I will choose 12vac, 15vac or 18vac (all *2 models), among these three.

If your transformer voltage are rated at load ,you can possibly go 12VAC to cover 12V and 13.5V, 15VAC to cover 15VDC .

[Pars]

On 5/28/2016 at 0:30 AM, mwl168 said:

I have a question about the voltage setting; if I want to configure the GRLV to output, say, 5.5VDC. Is it as simple as using a LT1021-5 5V voltage reference and change the R7 and R8 resistors to arrive at the 5.5VDC output?

My plan is to use the positive half of the GRLV to power the digital section of my DAC.  

 

On 5/28/2016 at 7:04 AM, kevin gilmore said:

never tried it that way.  a little over the top to use this to power digital.

does look like it will work.

 

 

On 5/28/2016 at 8:12 AM, kevin gilmore said:

its fine for digital, just seems a bit of overkill when a standard 3 terminal regulator is enough.

definitely change d4 to 5.1v

4.7v is too low and it will not start.

 

 

Some posts regarding lower voltages. I would think that the 10V LT1021 reference and 9.1V zeners should be OK for the 12/15 and definitely the 18V you were thinking about, but don't know for sure on the 12V. BTW, you can buy the LT1021 directly from Linear at a lower price, but only worth it if you are buying several.

When I was testing mine with a 24-26V transformer on a variac, it was regulating at 20V output down to around 18Vac input, a bit lower (17.9Vac). I wound up using the Antec 20V 520 transformer and it is working fine, and running at around 44C on the pass transistors powering an SS dynalo.

Edited December 26, 2016 by Pars

[kevin gilmore]

for 12V you might want to go with the 7 volt version of the reference and recalculate the resistors. And probably lower the 9.1v zener to 7V

I have not tested any of this.

15V should be fine with the standard parts

[mwl168]

16 minutes ago, kevin gilmore said:

for 12V you might want to go with the 7 volt version of the reference and recalculate the resistors. And probably lower the 9.1v zener to 7V

I have not tested any of this.

15V should be fine with the standard parts

I can confirm standard parts are fine with 15VDC. Have built one and has been running fine for months. 

[Laowei]

4 hours ago, mwl168 said:

I can confirm standard parts are fine with 15VDC. Have built one and has been running fine for months. 

Same here. 

[mtoc]

Update: suddenly I found a 2 x 18 VAC transformer in my house (which came from a power amp), so I don't need to buy it.

fisrtly I need the 13.5 VDC, so I just change the R8 and R9, then everything is fine? (R7 = R10 = 1.5k, plus the 10V version LT1021-10, 9.1V zeners)

R8 = R9

13.5 x 150 -1500 = R8 = 525

(R8 + R7) / R7 x 10 = 13.5

If using the 7V version
(R8 + R7) / R7 x 7 = 13.5
(R8 + R7) / R7 = 1.9286...
We change R7 to 2k to make it simpler
R8 = 1858 approximately

PS, good to hear this PSU running fine at 15V

[joehpj]

And you'll get two super hot BJT.

18v is way too high.

Edited December 27, 2016 by joehpj

[mtoc]

folks, gr lv dual version is right here

I've thrown all my extra dough on this build (including the dough for updating my old-granny-level-pc. been saved for a very very long time), so it better be good. Alth, I always fear that, my $1,350 is gonna get into smoke, and boom, boom, boom

pray for myself

[Pars]

You might be better off sticking with the 10V ref and using R7 of 700r (715 is closest e96 value) for 13.57Vdc. That is with a 2K R8. Only if Kevin thinks the 10V Vref will work fine for 13.5V.
What are you powering that needs 13.5Vdc? Seems like an odd value...


Sent from my iPhone using Tapatalk