Forums

Re-calibrating bench PSU

Started by bitrex July 17, 2020
I need another PSU on my bench and I remembered I had this
thing in my junk closet, a SMPS-based supply from www.mpje.com.

<https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/>

Something was wrong with it I couldn't remember but all I recalled was 
this is what the power PCB looked like, from the "factory" mind you:

<https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0>

sweet baby Jesus I don't know what they did to it. Looks like a return 
someone hacked back together in a shed in China. I plugged it in again 
and the NTC immediately smoked and cracked.

I cleaned up the board real good, retouched the solder joints and 
replaced the diodes (smaller TO-220, MURF1620CTG) with the same part and 
the switch (2SK1358) with a STW9NK90Z: 
<https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with 
an appropriate part.

Seems to work OK now and looks a lot better inside. Only problem is the 
display is off-calibration, and I doubt I'll ever find a service manual 
for the thing. On the controller PCB there's a button which seems to 
reset the current display to zero, it reads about 150 mA on all voltage 
settings with nothing connected otherwise. There are also two smt pots 
labeled "RP1" and "RP2". RP1 seems to adjust the voltage display value, 
and RP2 does...not very much as far as I can tell. They seem to interact 
with each other somewhat.

Anyone got any ideas on what combination of voltage, and current limit 
settings, load, and pot tweaks I should try to calibrate this thing?


On 7/17/2020 5:59 PM, bitrex wrote:
> I need another PSU on my bench and I remembered I had this > thing in my junk closet, a SMPS-based supply from www.mpje.com. > > <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> > > > Something was wrong with it I couldn't remember but all I recalled was > this is what the power PCB looked like, from the "factory" mind you: > > <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> > > sweet baby Jesus I don't know what they did to it. Looks like a return > someone hacked back together in a shed in China. I plugged it in again > and the NTC immediately smoked and cracked. > > I cleaned up the board real good, retouched the solder joints and > replaced the diodes (smaller TO-220, MURF1620CTG) with the same part and > the switch (2SK1358) with a STW9NK90Z: > <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with > an appropriate part. > > Seems to work OK now and looks a lot better inside. Only problem is the > display is off-calibration, and I doubt I'll ever find a service manual > for the thing. On the controller PCB there's a button which seems to > reset the current display to zero, it reads about 150 mA on all voltage > settings with nothing connected otherwise. There are also two smt pots > labeled "RP1" and "RP2". RP1 seems to adjust the voltage display value, > and RP2 does...not very much as far as I can tell. They seem to interact > with each other somewhat. > > Anyone got any ideas on what combination of voltage, and current limit > settings, load, and pot tweaks I should try to calibrate this thing? > >
Here's a schematic of the display/current and voltage sense logic board: <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image>
On Friday, July 17, 2020 at 2:59:31 PM UTC-7, bitrex wrote:
> I need another PSU on my bench and I remembered I had this > thing in my junk closet, a SMPS-based supply from www.mpje.com. > > <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> > > Something was wrong with it I couldn't remember but all I recalled was > this is what the power PCB looked like, from the "factory" mind you: > > <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> > > sweet baby Jesus I don't know what they did to it. Looks like a return > someone hacked back together in a shed in China. I plugged it in again > and the NTC immediately smoked and cracked. > > I cleaned up the board real good, retouched the solder joints and > replaced the diodes (smaller TO-220, MURF1620CTG) with the same part and > the switch (2SK1358) with a STW9NK90Z: > <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with > an appropriate part. > > Seems to work OK now and looks a lot better inside. Only problem is the > display is off-calibration, and I doubt I'll ever find a service manual > for the thing. On the controller PCB there's a button which seems to > reset the current display to zero, it reads about 150 mA on all voltage > settings with nothing connected otherwise. There are also two smt pots > labeled "RP1" and "RP2". RP1 seems to adjust the voltage display value, > and RP2 does...not very much as far as I can tell. They seem to interact > with each other somewhat. > > Anyone got any ideas on what combination of voltage, and current limit > settings, load, and pot tweaks I should try to calibrate this thing?
My guess is the second trimpot is used to calibrate the current limit knob. Try applying a short to the output through an ammeter and see if turning the trimpot changes the display.
On 7/17/2020 10:22 PM, bitrex wrote:
> On 7/17/2020 5:59 PM, bitrex wrote: >> I need another PSU on my bench and I remembered I had this >> thing in my junk closet, a SMPS-based supply from www.mpje.com. >> >> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >> >> >> Something was wrong with it I couldn't remember but all I recalled was >> this is what the power PCB looked like, from the "factory" mind you: >> >> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >> >> >> sweet baby Jesus I don't know what they did to it. Looks like a return >> someone hacked back together in a shed in China. I plugged it in again >> and the NTC immediately smoked and cracked. >> >> I cleaned up the board real good, retouched the solder joints and >> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part >> and the switch (2SK1358) with a STW9NK90Z: >> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >> an appropriate part. >> >> Seems to work OK now and looks a lot better inside. Only problem is >> the display is off-calibration, and I doubt I'll ever find a service >> manual for the thing. On the controller PCB there's a button which >> seems to reset the current display to zero, it reads about 150 mA on >> all voltage settings with nothing connected otherwise. There are also >> two smt pots labeled "RP1" and "RP2". RP1 seems to adjust the voltage >> display value, and RP2 does...not very much as far as I can tell. They >> seem to interact with each other somewhat. >> >> Anyone got any ideas on what combination of voltage, and current limit >> settings, load, and pot tweaks I should try to calibrate this thing? >> >> > > Here's a schematic of the display/current and voltage sense logic board: > > <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> >
I can wrangle current sense trimpot "RV1" for days and the voltage across C10 never changes so shorted pot, looks like. :(
On 7/17/2020 10:56 PM, sea moss wrote:
> On Friday, July 17, 2020 at 2:59:31 PM UTC-7, bitrex wrote: >> I need another PSU on my bench and I remembered I had this >> thing in my junk closet, a SMPS-based supply from www.mpje.com. >> >> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >> >> Something was wrong with it I couldn't remember but all I recalled was >> this is what the power PCB looked like, from the "factory" mind you: >> >> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >> >> sweet baby Jesus I don't know what they did to it. Looks like a return >> someone hacked back together in a shed in China. I plugged it in again >> and the NTC immediately smoked and cracked. >> >> I cleaned up the board real good, retouched the solder joints and >> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part and >> the switch (2SK1358) with a STW9NK90Z: >> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >> an appropriate part. >> >> Seems to work OK now and looks a lot better inside. Only problem is the >> display is off-calibration, and I doubt I'll ever find a service manual >> for the thing. On the controller PCB there's a button which seems to >> reset the current display to zero, it reads about 150 mA on all voltage >> settings with nothing connected otherwise. There are also two smt pots >> labeled "RP1" and "RP2". RP1 seems to adjust the voltage display value, >> and RP2 does...not very much as far as I can tell. They seem to interact >> with each other somewhat. >> >> Anyone got any ideas on what combination of voltage, and current limit >> settings, load, and pot tweaks I should try to calibrate this thing? > > My guess is the second trimpot is used to calibrate the current limit knob. Try applying a short to the output through an ammeter and see if turning the trimpot changes the display. >
Yep that is right. Schem: <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> Display reads the correct voltage but wrong current. The voltage across current sense output cap C10 never budges much no matter how long I spin that trimpot so looks like I've also got a shorted pot, there. There are two more pots on the control board to adjust the current and output voltage limits of the knobs so I guess you could "mod" this into a higher voltage supply straightforwardly, so long as you stayed within the limits of the associated components and tweaked the current limit down to keep max output power limit the same.
On Friday, July 17, 2020 at 7:58:36 PM UTC-7, bitrex wrote:
> On 7/17/2020 10:22 PM, bitrex wrote: > > On 7/17/2020 5:59 PM, bitrex wrote: > >> I need another PSU on my bench and I remembered I had this > >> thing in my junk closet, a SMPS-based supply from www.mpje.com. > >> > >> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> > >> > >> > >> Something was wrong with it I couldn't remember but all I recalled was > >> this is what the power PCB looked like, from the "factory" mind you: > >> > >> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> > >> > >> > >> sweet baby Jesus I don't know what they did to it. Looks like a return > >> someone hacked back together in a shed in China. I plugged it in again > >> and the NTC immediately smoked and cracked. > >> > >> I cleaned up the board real good, retouched the solder joints and > >> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part > >> and the switch (2SK1358) with a STW9NK90Z: > >> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with > >> an appropriate part. > >> > >> Seems to work OK now and looks a lot better inside. Only problem is > >> the display is off-calibration, and I doubt I'll ever find a service > >> manual for the thing. On the controller PCB there's a button which > >> seems to reset the current display to zero, it reads about 150 mA on > >> all voltage settings with nothing connected otherwise. There are also > >> two smt pots labeled "RP1" and "RP2". RP1 seems to adjust the voltage > >> display value, and RP2 does...not very much as far as I can tell. They > >> seem to interact with each other somewhat. > >> > >> Anyone got any ideas on what combination of voltage, and current limit > >> settings, load, and pot tweaks I should try to calibrate this thing? > >> > >> > > > > Here's a schematic of the display/current and voltage sense logic board: > > > > <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> > > > > I can wrangle current sense trimpot "RV1" for days and the voltage > across C10 never changes so shorted pot, looks like. :(
Does the trimpot have a mechanical start and stop point, beyond which you can't turn it? Some of them will allow you to keep turning forever, and once it hits the end of the resistance range the value doesn't change anymore as you continue to turn. So you might have to back up by like 20 revolutions to find the active range.
On 7/17/2020 11:03 PM, bitrex wrote:
> On 7/17/2020 10:56 PM, sea moss wrote: >> On Friday, July 17, 2020 at 2:59:31 PM UTC-7, bitrex wrote: >>> I need another PSU on my bench and I remembered I had this >>> thing in my junk closet, a SMPS-based supply from www.mpje.com. >>> >>> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >>> >>> >>> Something was wrong with it I couldn't remember but all I recalled was >>> this is what the power PCB looked like, from the "factory" mind you: >>> >>> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >>> >>> >>> sweet baby Jesus I don't know what they did to it. Looks like a return >>> someone hacked back together in a shed in China. I plugged it in again >>> and the NTC immediately smoked and cracked. >>> >>> I cleaned up the board real good, retouched the solder joints and >>> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part and >>> the switch (2SK1358) with a STW9NK90Z: >>> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >>> an appropriate part. >>> >>> Seems to work OK now and looks a lot better inside. Only problem is the >>> display is off-calibration, and I doubt I'll ever find a service manual >>> for the thing. On the controller PCB there's a button which seems to >>> reset the current display to zero, it reads about 150 mA on all voltage >>> settings with nothing connected otherwise. There are also two smt pots >>> labeled "RP1" and "RP2". RP1 seems to adjust the voltage display value, >>> and RP2 does...not very much as far as I can tell. They seem to interact >>> with each other somewhat. >>> >>> Anyone got any ideas on what combination of voltage, and current limit >>> settings, load, and pot tweaks I should try to calibrate this thing? >> >> My guess is the second trimpot is used to calibrate the current limit >> knob.&nbsp; Try applying a short to the output through an ammeter and see >> if turning the trimpot changes the display. >> > > Yep that is right. Schem: > > <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> > > > Display reads the correct voltage but wrong current. The voltage across > current sense output cap C10 never budges much no matter how long I spin > that trimpot so looks like I've also got a shorted pot, there. > > There are two more pots on the control board to adjust the current and > output voltage limits of the knobs so I guess you could "mod" this into > a higher voltage supply straightforwardly, so long as you stayed within > the limits of the associated components and tweaked the current limit > down to keep max output power limit the same. >
They're not using low-offset op amps or precision resistors here as far as I can tell so getting the display readout as good as it can be is I think one of those things like intonating a guitar, tweak it at the low end, turn the voltage up to max and it'll read high there, tweak it down a bit, turn it back down to the low end and adjust it again there, and just kinda do the best you can
On 7/17/2020 11:05 PM, sea moss wrote:
> On Friday, July 17, 2020 at 7:58:36 PM UTC-7, bitrex wrote: >> On 7/17/2020 10:22 PM, bitrex wrote: >>> On 7/17/2020 5:59 PM, bitrex wrote: >>>> I need another PSU on my bench and I remembered I had this >>>> thing in my junk closet, a SMPS-based supply from www.mpje.com. >>>> >>>> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >>>> >>>> >>>> Something was wrong with it I couldn't remember but all I recalled was >>>> this is what the power PCB looked like, from the "factory" mind you: >>>> >>>> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >>>> >>>> >>>> sweet baby Jesus I don't know what they did to it. Looks like a return >>>> someone hacked back together in a shed in China. I plugged it in again >>>> and the NTC immediately smoked and cracked. >>>> >>>> I cleaned up the board real good, retouched the solder joints and >>>> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part >>>> and the switch (2SK1358) with a STW9NK90Z: >>>> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >>>> an appropriate part. >>>> >>>> Seems to work OK now and looks a lot better inside. Only problem is >>>> the display is off-calibration, and I doubt I'll ever find a service >>>> manual for the thing. On the controller PCB there's a button which >>>> seems to reset the current display to zero, it reads about 150 mA on >>>> all voltage settings with nothing connected otherwise. There are also >>>> two smt pots labeled "RP1" and "RP2". RP1 seems to adjust the voltage >>>> display value, and RP2 does...not very much as far as I can tell. They >>>> seem to interact with each other somewhat. >>>> >>>> Anyone got any ideas on what combination of voltage, and current limit >>>> settings, load, and pot tweaks I should try to calibrate this thing? >>>> >>>> >>> >>> Here's a schematic of the display/current and voltage sense logic board: >>> >>> <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> >>> >> >> I can wrangle current sense trimpot "RV1" for days and the voltage >> across C10 never changes so shorted pot, looks like. :( > > Does the trimpot have a mechanical start and stop point, beyond which you can't turn it? Some of them will allow you to keep turning forever, and once it hits the end of the resistance range the value doesn't change anymore as you continue to turn. So you might have to back up by like 20 revolutions to find the active range. >
No mechanical limits, but yeah I should check if I haven't just spun it out of the active range by accident
On 7/17/2020 11:05 PM, sea moss wrote:
> On Friday, July 17, 2020 at 7:58:36 PM UTC-7, bitrex wrote: >> On 7/17/2020 10:22 PM, bitrex wrote: >>> On 7/17/2020 5:59 PM, bitrex wrote: >>>> I need another PSU on my bench and I remembered I had this >>>> thing in my junk closet, a SMPS-based supply from www.mpje.com. >>>> >>>> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >>>> >>>> >>>> Something was wrong with it I couldn't remember but all I recalled was >>>> this is what the power PCB looked like, from the "factory" mind you: >>>> >>>> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >>>> >>>> >>>> sweet baby Jesus I don't know what they did to it. Looks like a return >>>> someone hacked back together in a shed in China. I plugged it in again >>>> and the NTC immediately smoked and cracked. >>>> >>>> I cleaned up the board real good, retouched the solder joints and >>>> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part >>>> and the switch (2SK1358) with a STW9NK90Z: >>>> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >>>> an appropriate part. >>>> >>>> Seems to work OK now and looks a lot better inside. Only problem is >>>> the display is off-calibration, and I doubt I'll ever find a service >>>> manual for the thing. On the controller PCB there's a button which >>>> seems to reset the current display to zero, it reads about 150 mA on >>>> all voltage settings with nothing connected otherwise. There are also >>>> two smt pots labeled "RP1" and "RP2". RP1 seems to adjust the voltage >>>> display value, and RP2 does...not very much as far as I can tell. They >>>> seem to interact with each other somewhat. >>>> >>>> Anyone got any ideas on what combination of voltage, and current limit >>>> settings, load, and pot tweaks I should try to calibrate this thing? >>>> >>>> >>> >>> Here's a schematic of the display/current and voltage sense logic board: >>> >>> <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> >>> >> >> I can wrangle current sense trimpot "RV1" for days and the voltage >> across C10 never changes so shorted pot, looks like. :( > > Does the trimpot have a mechanical start and stop point, beyond which you can't turn it? Some of them will allow you to keep turning forever, and once it hits the end of the resistance range the value doesn't change anymore as you continue to turn. So you might have to back up by like 20 revolutions to find the active range. >
Meh, can't find anywhere on it "active." Retouched the solder joints that didn't help. Have to pull it and see what's what.
On 7/17/2020 11:05 PM, sea moss wrote:
> On Friday, July 17, 2020 at 7:58:36 PM UTC-7, bitrex wrote: >> On 7/17/2020 10:22 PM, bitrex wrote: >>> On 7/17/2020 5:59 PM, bitrex wrote: >>>> I need another PSU on my bench and I remembered I had this >>>> thing in my junk closet, a SMPS-based supply from www.mpje.com. >>>> >>>> <https://www.summet.com/blog/2012/01/22/review-mpja-9615-0-30-volt-0-3-amp-bench-power-supply/> >>>> >>>> >>>> Something was wrong with it I couldn't remember but all I recalled was >>>> this is what the power PCB looked like, from the "factory" mind you: >>>> >>>> <https://www.dropbox.com/s/65km6pbq8wzch54/IMG_20171005_104357775.jpg?dl=0> >>>> >>>> >>>> sweet baby Jesus I don't know what they did to it. Looks like a return >>>> someone hacked back together in a shed in China. I plugged it in again >>>> and the NTC immediately smoked and cracked. >>>> >>>> I cleaned up the board real good, retouched the solder joints and >>>> replaced the diodes (smaller TO-220, MURF1620CTG) with the same part >>>> and the switch (2SK1358) with a STW9NK90Z: >>>> <https://www.mouser.com/ProductDetail/511-STW9NK90Z>, and the NTC with >>>> an appropriate part. >>>> >>>> Seems to work OK now and looks a lot better inside. Only problem is >>>> the display is off-calibration, and I doubt I'll ever find a service >>>> manual for the thing. On the controller PCB there's a button which >>>> seems to reset the current display to zero, it reads about 150 mA on >>>> all voltage settings with nothing connected otherwise. There are also >>>> two smt pots labeled "RP1" and "RP2". RP1 seems to adjust the voltage >>>> display value, and RP2 does...not very much as far as I can tell. They >>>> seem to interact with each other somewhat. >>>> >>>> Anyone got any ideas on what combination of voltage, and current limit >>>> settings, load, and pot tweaks I should try to calibrate this thing? >>>> >>>> >>> >>> Here's a schematic of the display/current and voltage sense logic board: >>> >>> <https://www.eevblog.com/forum/projects/cheap-benchtop-psu-upgrade/?action=dlattach;attach=398874;image> >>> >> >> I can wrangle current sense trimpot "RV1" for days and the voltage >> across C10 never changes so shorted pot, looks like. :( > > Does the trimpot have a mechanical start and stop point, beyond which you can't turn it? Some of them will allow you to keep turning forever, and once it hits the end of the resistance range the value doesn't change anymore as you continue to turn. So you might have to back up by like 20 revolutions to find the active range. >
Intermittent pot, looks like. Resistance drops to 1.5 ohms as you turn it at various points. I'll try bodging in a better 1k trimmer this and weekend see how it do.