Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.

Here's the schematic (PDF), logic board only, unfortunately but the
power board seems to be working OK.

https://www.eevblog.com/forum/testgear/tekpower-tp3005t-variable-linear-dc-power-supply/?action=dlattach;attach=249613

Overcurrent detection seems to work by voltages IS- and IS+ coming in
from the off-board current shunt and sent to pin 19 of the PIC, which I
believe is a comparator-interrupt pin, and that seems to work ok, and
the voltages out of that op-amp and across the current shunt look OK.

Constant current mode seems to work by using TL082 section "N18BB" as a
comparator open-loop, and then the loop is closed around back from the
ADJ line that goes off-board to the power regulator board, and the
current-set DAC output from the PIC, comparing the two inputs against
each other.

Constant-voltage mode works similarly the voltage-set DAC value from the
PIC is compared against the voltage at the output and a line is asserted
when they match. The output voltage in CV mode looks OK.

It's not clear to me though how what the current actually is, in
constant-voltage mode, is sensed by the PIC, as I don't immediately see
a pin on it where the current-shunt voltage is being read by the uP, so
it's hard for me to know where to look for the problem as nothing
obvious is amiss, and continuity on all the connections between the
boards to the PCBs seems OK.

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.

Here's the schematic (PDF), logic board only, unfortunately but the
power board seems to be working OK.

https://www.eevblog.com/forum/testgear/tekpower-tp3005t-variable-linear-dc-power-supply/?action=dlattach;attach=249613


Overcurrent detection seems to work by voltages IS- and IS+ coming in
from the off-board current shunt and sent to pin 19 of the PIC, which I
believe is a comparator-interrupt pin, and that seems to work ok, and
the voltages out of that op-amp and across the current shunt look OK.

Constant current mode seems to work by using TL082 section "N18BB" as a
comparator open-loop, and then the loop is closed around back from the
ADJ line that goes off-board to the power regulator board, and the
current-set DAC output from the PIC, comparing the two inputs against
each other.

Constant-voltage mode works similarly the voltage-set DAC value from the
PIC is compared against the voltage at the output and a line is asserted
when they match. The output voltage in CV mode looks OK.

It's not clear to me though how what the current actually is, in
constant-voltage mode, is sensed by the PIC, as I don't immediately see
a pin on it where the current-shunt voltage is being read by the uP, so
it's hard for me to know where to look for the problem as nothing
obvious is amiss, and continuity on all the connections between the
boards to the PCBs seems OK.



Only other thing I notice is that the logic board ground and the PSU
positive output seem to be connected, with the logic board fed by a
separate isolated tap on the transformer. And it looks like one of the
ADC inputs on the PIC is also connected to the logic board ground, so
maybe some kind of differential ADC arrangement but how to diagnose
still eludes me.

On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL

On 8/11/2020 8:36 AM, legg wrote:
On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL


Yeah the diagnostic process is definitely not helped by the sloppy and
confusing schematic. Before I jump into op-amp swapping I'm trying to
understand how the current-sensing works in CV mode and is read by the
processor so maybe I can trace it back, but I'm not very familiar with
the PIC. Maybe it's reading the value from the "CCP" pin somehow that
"VIS" connects to, pin 19, but the impression I got from the datasheet
is those are some kind of comparator inputs for the
over-current/over-voltage sensing, not ADCs. Not sure.

Like a lot of gear from China at this price point it's designed to meet
a budget, but not badly designed to meet a budget and otherwise neatly
constructed, but then often let down by some sketchy soldering on the
connectors and through-hole parts. Often can "fix" this stuff by tapping
the board to find the intermittent joint or just re-touching all the
connectors but my first round of that didn't work, this time.

On Tue, 11 Aug 2020 12:12:46 -0400, bitrex wrote:

On 8/11/2020 8:36 AM, legg wrote:
On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL


Yeah the diagnostic process is definitely not helped by the sloppy and
confusing schematic. Before I jump into op-amp swapping I'm trying to
understand how the current-sensing works in CV mode and is read by the
processor so maybe I can trace it back, but I'm not very familiar with
the PIC. Maybe it's reading the value from the "CCP" pin somehow that
"VIS" connects to, pin 19, but the impression I got from the datasheet
is those are some kind of comparator inputs for the
over-current/over-voltage sensing, not ADCs. Not sure.

Like a lot of gear from China at this price point it's designed to meet
a budget, but not badly designed to meet a budget and otherwise neatly
constructed, but then often let down by some sketchy soldering on the
connectors and through-hole parts. Often can "fix" this stuff by tapping
the board to find the intermittent joint or just re-touching all the
connectors but my first round of that didn't work, this time.

You can always measure IC output pin voltages during the
discontinuity, before pulling them.

In a close-packed assembly, this isn't always much fun - but once
exposed, they're easier to replace anyways.

V13 is just digital.

RL

On 8/11/2020 3:38 PM, legg wrote:
On Tue, 11 Aug 2020 12:12:46 -0400, bitrex wrote:

On 8/11/2020 8:36 AM, legg wrote:
On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL


Yeah the diagnostic process is definitely not helped by the sloppy and
confusing schematic. Before I jump into op-amp swapping I'm trying to
understand how the current-sensing works in CV mode and is read by the
processor so maybe I can trace it back, but I'm not very familiar with
the PIC. Maybe it's reading the value from the "CCP" pin somehow that
"VIS" connects to, pin 19, but the impression I got from the datasheet
is those are some kind of comparator inputs for the
over-current/over-voltage sensing, not ADCs. Not sure.

Like a lot of gear from China at this price point it's designed to meet
a budget, but not badly designed to meet a budget and otherwise neatly
constructed, but then often let down by some sketchy soldering on the
connectors and through-hole parts. Often can "fix" this stuff by tapping
the board to find the intermittent joint or just re-touching all the
connectors but my first round of that didn't work, this time.

You can always measure IC output pin voltages during the
discontinuity, before pulling them.

In a close-packed assembly, this isn't always much fun - but once
exposed, they're easier to replace anyways.

V13 is just digital.

RL


Still no luck on figuring out this one. Tested the current op amp
output with respect to the logic board ground, which is the output
positive terminal, to avoid doing large voltage divisions to feed the
3.3V uP at the higher voltage settings I'm guessing. Output looks okay,
and the constant current mode works correctly, it regulates down the
output voltage if the load goes over the current set point. so the "ADJ"
pin to the power board, transistor "V12" must be outputting signal OK.
so the current sense amp from the shunt and DAC and all that must be
able to enter closed-loop operation.

Still, the output indicator always reads 0 mA when in CV mode. Ugh.
Where does the current sense amp voltage actually enter the PIC via an
ADC? Can't find it anywhere. Maybe there's a processor fault or
something, unfortunately.

On Sat, 15 Aug 2020 01:38:58 -0400, bitrex wrote:

On 8/11/2020 3:38 PM, legg wrote:
On Tue, 11 Aug 2020 12:12:46 -0400, bitrex wrote:

On 8/11/2020 8:36 AM, legg wrote:
On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL


Yeah the diagnostic process is definitely not helped by the sloppy and
confusing schematic. Before I jump into op-amp swapping I'm trying to
understand how the current-sensing works in CV mode and is read by the
processor so maybe I can trace it back, but I'm not very familiar with
the PIC. Maybe it's reading the value from the "CCP" pin somehow that
"VIS" connects to, pin 19, but the impression I got from the datasheet
is those are some kind of comparator inputs for the
over-current/over-voltage sensing, not ADCs. Not sure.

Like a lot of gear from China at this price point it's designed to meet
a budget, but not badly designed to meet a budget and otherwise neatly
constructed, but then often let down by some sketchy soldering on the
connectors and through-hole parts. Often can "fix" this stuff by tapping
the board to find the intermittent joint or just re-touching all the
connectors but my first round of that didn't work, this time.

You can always measure IC output pin voltages during the
discontinuity, before pulling them.

In a close-packed assembly, this isn't always much fun - but once
exposed, they're easier to replace anyways.

V13 is just digital.

RL


Still no luck on figuring out this one. Tested the current op amp
output with respect to the logic board ground, which is the output
positive terminal, to avoid doing large voltage divisions to feed the
3.3V uP at the higher voltage settings I'm guessing. Output looks okay,
and the constant current mode works correctly, it regulates down the
output voltage if the load goes over the current set point. so the "ADJ"
pin to the power board, transistor "V12" must be outputting signal OK.
so the current sense amp from the shunt and DAC and all that must be
able to enter closed-loop operation.

Still, the output indicator always reads 0 mA when in CV mode. Ugh.
Where does the current sense amp voltage actually enter the PIC via an
ADC? Can't find it anywhere. Maybe there's a processor fault or
something, unfortunately.

DAC input current sense pin is VIS (pin19); so C22 short, R30 open, or
contact trace problems.

?

RL

On 8/15/2020 11:00 AM, legg wrote:
On Sat, 15 Aug 2020 01:38:58 -0400, bitrex wrote:

On 8/11/2020 3:38 PM, legg wrote:
On Tue, 11 Aug 2020 12:12:46 -0400, bitrex wrote:

On 8/11/2020 8:36 AM, legg wrote:
On Tue, 11 Aug 2020 03:54:01 -0400, bitrex wrote:

On 8/11/2020 3:41 AM, bitrex wrote:
Problem with my Tekpower PSU:

https://tekpower.us/downloadable/download/sample/sample_id/1/

Here's a teardown on YouTube:

https://youtu.be/Nfe28vGW6Ik

Problem is it's developed a fault where the current meter always reads 0
when in constant-voltage mode. Everything else seems to work fine and
it's putting out the right voltages, current limit works fine, when load
current exceeds the set point it current limits OK and shows the correct
value. Drop back and the display reads 0, again.
snip

Have you replaced the OP07?

Could be one output is unable to pull in one direction.

Or N18A . . .

Biasing to V13 doesn't look right, in the schematic, so
you may have to deal with some typos there.

RL


Yeah the diagnostic process is definitely not helped by the sloppy and
confusing schematic. Before I jump into op-amp swapping I'm trying to
understand how the current-sensing works in CV mode and is read by the
processor so maybe I can trace it back, but I'm not very familiar with
the PIC. Maybe it's reading the value from the "CCP" pin somehow that
"VIS" connects to, pin 19, but the impression I got from the datasheet
is those are some kind of comparator inputs for the
over-current/over-voltage sensing, not ADCs. Not sure.

Like a lot of gear from China at this price point it's designed to meet
a budget, but not badly designed to meet a budget and otherwise neatly
constructed, but then often let down by some sketchy soldering on the
connectors and through-hole parts. Often can "fix" this stuff by tapping
the board to find the intermittent joint or just re-touching all the
connectors but my first round of that didn't work, this time.

You can always measure IC output pin voltages during the
discontinuity, before pulling them.

In a close-packed assembly, this isn't always much fun - but once
exposed, they're easier to replace anyways.

V13 is just digital.

RL


Still no luck on figuring out this one. Tested the current op amp
output with respect to the logic board ground, which is the output
positive terminal, to avoid doing large voltage divisions to feed the
3.3V uP at the higher voltage settings I'm guessing. Output looks okay,
and the constant current mode works correctly, it regulates down the
output voltage if the load goes over the current set point. so the "ADJ"
pin to the power board, transistor "V12" must be outputting signal OK.
so the current sense amp from the shunt and DAC and all that must be
able to enter closed-loop operation.

Still, the output indicator always reads 0 mA when in CV mode. Ugh.
Where does the current sense amp voltage actually enter the PIC via an
ADC? Can't find it anywhere. Maybe there's a processor fault or
something, unfortunately.

DAC input current sense pin is VIS (pin19); so C22 short, R30 open, or
contact trace problems.

?

RL


Thanks legg, I'll check those.