Samles SSW2000 inverter early limit...

L

legg

Guest
I\'ve got an SSW2000 repair that seems to be turning into a science
project.

The customer bought it cheap off Kijiji. It didn\'t \'go\'.

I found all 10 of the 30A automotive fuses blown, suggestting that
the last thing done to it was a reversed battery connection.
None of the DC inverter fets were damaged - This section proved to
run when a normal downsteam load was present.

The two main HV bulk caps were dried out and had vented oton the
wrapped heatsink. This over a long enough time period that the
vented material had collected air fluff as it dried.A small spot
of vented material had been deflected onto the LV driver control
board, bridging board vias and requiring scrubbing w alcohol to
remove.

One side of the inverter bridge was shorted through the switching
fets. When replaced, the recapped inverter ran normally under
lighter loads (below 500W).

I might mention that one of the unit\'s fans was dead, as well.

There is a small visible nonlinearity in the negative-going
sine output \'increasing\' edge, which looks increasingly like
undamped response as load increases. This (negative generating
pwm) was the side of the bridge that had failed previously.

At some point of increasing load, the negative going sine
output flat-tops - coincident with the positive going side
cutting out completely on its falling edge, then the unit
latches off to an overload limit.

I\'m not sure if this performance is aggravated by the loading
method - resistive through a variac, so any DC imbalance can
result in largish mag current imbalance.

Input to the low frequency AC limiting circuit seems normal,
though it is designed with a ridiculously high impedance -
a multimegohm RC filter from a 70milliohm sense resistor -
possibly to cut down on C size. It floats on the swinging
polarity of the negative bulk caps. Even some of my go-to
diff probes leak enough to make this a \'hands-off\' test point
while in operation, but what I can see of larger signal stuff
at 50:1 diff looks normal.

There are no schematics available for the 3x control cct boards,
which are basically 339s and 324s working with schmitt triggers
and flip flops. Gate drive is optoisolated x4 with DC restored
bootstrap power on each. LF AC reg (tl431)and fault signals are
optoisolated

A main board schematic is available on groups.io in Electronics101.
This shows that any control cctry components on the main board
are basically concerned with fan control or external comms.
Red detail indicates modifications not present on production
(this) units - again mainly fan control.

http://ve3ute.ca/query/Samlex_SSW-2000_main_board_schema.pdf

Any ideas or similar experience out there?

RL

Housekeeping supplies are reverse polarity protected, simple
and solid. AC voltage output LF amplitude is well regulated.
 
On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

Specs:
<https://samlexamerica.com/wp-content/uploads/2019/12/12002-SSW-2000-12A-0117.pdf>

User manual, which includes some troubleshooting hints.
<https://samlexamerica.com/wp-content/uploads/2019/12/11002-SSW-350-600-1000-1500-2000-12A-0117-Lrez.pdf>

YouTube videos showing SSW-2000 repair (5 parts)
<https://www.youtube.com/watch?v=Vq-8cV2Y8Ac>
<https://www.youtube.com/watch?v=GMvSP4ojZRo>
<https://www.youtube.com/watch?v=eIwK7HHqyhw>
<https://www.youtube.com/watch?v=uJRR6NKsT3U>
<https://www.youtube.com/watch?v=s0mxwoZOqlQ>

Text accompanying Part 5:

\"Samlex provided me with the main PCB schematics, and a new updated
input drive circuitry daughter board. It was an updated design and
more reliable. I changed out all the input FETs with some quality
International Rectifier brand, and input caps to Nichicon brand with a
higher voltage rating. A few drive transistors burnt/shorted and
replaced those.\"

Maybe give Samlex a call and ask for a better schematic?


--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558
 
On Sun, 06 Nov 2022 10:07:50 -0800, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

Specs:
https://samlexamerica.com/wp-content/uploads/2019/12/12002-SSW-2000-12A-0117.pdf

User manual, which includes some troubleshooting hints.
https://samlexamerica.com/wp-content/uploads/2019/12/11002-SSW-350-600-1000-1500-2000-12A-0117-Lrez.pdf

YouTube videos showing SSW-2000 repair (5 parts)
https://www.youtube.com/watch?v=Vq-8cV2Y8Ac
https://www.youtube.com/watch?v=GMvSP4ojZRo
https://www.youtube.com/watch?v=eIwK7HHqyhw
https://www.youtube.com/watch?v=uJRR6NKsT3U
https://www.youtube.com/watch?v=s0mxwoZOqlQ

Text accompanying Part 5:

\"Samlex provided me with the main PCB schematics, and a new updated
input drive circuitry daughter board. It was an updated design and
more reliable. I changed out all the input FETs with some quality
International Rectifier brand, and input caps to Nichicon brand with a
higher voltage rating. A few drive transistors burnt/shorted and
replaced those.\"

Maybe give Samlex a call and ask for a better schematic?

Samlex would not release schematics or service info.

I have read most of the available manuals and can assure
you that this isn\'t a battery supply issue.

I saw these youtube videos, and low limiting wasn\'t
one of the issues the guy tackled. He coaxed a dead
driver board out of Samlex to get the thing going,
finally. Not sure he loaded it above a KW.

He did a lot of random recapping and replaced a fan driver
transistor.

I can\'t see anything here below 400W load that would be
an issue above that power level. I typoed the AC current
sense value - ~15mR (seven resistors in parallel causing
me to expectorate the 70mR figure).

Have measured and tack-on substituted most of the
power filter film caps without joy.

Will be doing things like swapping over parts from one
phase to the other, to see if symptoms follow, but
this will involve some serious (de)soldering on
light weight copper boards.

Will need to cut power copper traces to check
currents - again, more science project than repair.

Ugh.

RL
 
On Sun, 06 Nov 2022 10:07:50 -0800, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

Specs:
https://samlexamerica.com/wp-content/uploads/2019/12/12002-SSW-2000-12A-0117.pdf

User manual, which includes some troubleshooting hints.
https://samlexamerica.com/wp-content/uploads/2019/12/11002-SSW-350-600-1000-1500-2000-12A-0117-Lrez.pdf

YouTube videos showing SSW-2000 repair (5 parts)
https://www.youtube.com/watch?v=Vq-8cV2Y8Ac
https://www.youtube.com/watch?v=GMvSP4ojZRo
https://www.youtube.com/watch?v=eIwK7HHqyhw
https://www.youtube.com/watch?v=uJRR6NKsT3U
https://www.youtube.com/watch?v=s0mxwoZOqlQ

Text accompanying Part 5:

\"Samlex provided me with the main PCB schematics, and a new updated
input drive circuitry daughter board. It was an updated design and
more reliable. I changed out all the input FETs with some quality
International Rectifier brand, and input caps to Nichicon brand with a
higher voltage rating. A few drive transistors burnt/shorted and
replaced those.\"

Maybe give Samlex a call and ask for a better schematic?

I\'ve cut tracks to monitor power fet current in the offending
side of the bridge, and see noticeable crossconduction there
under all load conditions, when pwming.

The fet driver board, powered and running in isolation, shows
no gate drive issues - dead time is actually generated on this
board. ~500nS should be sufficient with 10R series gate
resistors, and there is no difficulty demonstrated in driving
capacitive loads.

A modest negative \'off\' bias is developed by zener/dc restoration
on a series coupling capacitor.

It\'s basically just an isolated buffer with full fet control
deveoped in a separate small signal section.

The presence of cross-conduction makes no sense, if the gate
drive circuit is testing functional, though it would easily be
attributed as causal in early limiting of the phase arm.

RL
 
On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

The customer bought it cheap off Kijiji. It didn\'t \'go\'.

I found all 10 of the 30A automotive fuses blown, suggestting that
the last thing done to it was a reversed battery connection.
None of the DC inverter fets were damaged - This section proved to
run when a normal downsteam load was present.

The two main HV bulk caps were dried out and had vented oton the
wrapped heatsink. This over a long enough time period that the
vented material had collected air fluff as it dried.A small spot
of vented material had been deflected onto the LV driver control
board, bridging board vias and requiring scrubbing w alcohol to
remove.

One side of the inverter bridge was shorted through the switching
fets. When replaced, the recapped inverter ran normally under
lighter loads (below 500W).

I might mention that one of the unit\'s fans was dead, as well.

There is a small visible nonlinearity in the negative-going
sine output \'increasing\' edge, which looks increasingly like
undamped response as load increases. This (negative generating
pwm) was the side of the bridge that had failed previously.

At some point of increasing load, the negative going sine
output flat-tops - coincident with the positive going side
cutting out completely on its falling edge, then the unit
latches off to an overload limit.

I\'m not sure if this performance is aggravated by the loading
method - resistive through a variac, so any DC imbalance can
result in largish mag current imbalance.

Input to the low frequency AC limiting circuit seems normal,
though it is designed with a ridiculously high impedance -
a multimegohm RC filter from a 70milliohm sense resistor -
possibly to cut down on C size. It floats on the swinging
polarity of the negative bulk caps. Even some of my go-to
diff probes leak enough to make this a \'hands-off\' test point
while in operation, but what I can see of larger signal stuff
at 50:1 diff looks normal.

There are no schematics available for the 3x control cct boards,
which are basically 339s and 324s working with schmitt triggers
and flip flops. Gate drive is optoisolated x4 with DC restored
bootstrap power on each. LF AC reg (tl431)and fault signals are
optoisolated

A main board schematic is available on groups.io in Electronics101.
This shows that any control cctry components on the main board
are basically concerned with fan control or external comms.
Red detail indicates modifications not present on production
(this) units - again mainly fan control.

http://ve3ute.ca/query/Samlex_SSW-2000_main_board_schema.pdf

Any ideas or similar experience out there?

RL

Housekeeping supplies are reverse polarity protected, simple
and solid. AC voltage output LF amplitude is well regulated.

A solution is found accidentally - could get >1KW out of this
when a diff probe was attached between the AC vreg divider
point and the floating reference ground for that section.
That\'s between the junctions of R27/R28/R56 and the \'bottom\'
of VR1.

This is the rectified but unfiltered AC output, to provide feedback
for 60Hz voltage regulation.

Why? Who the hell knows. I tacked on a 100pf capacitor, burned the
thing in and booted it out the door.

RL
 
On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

The customer bought it cheap off Kijiji. It didn\'t \'go\'.

I found all 10 of the 30A automotive fuses blown, suggestting that
the last thing done to it was a reversed battery connection.
None of the DC inverter fets were damaged - This section proved to
run when a normal downsteam load was present.

The two main HV bulk caps were dried out and had vented oton the
wrapped heatsink. This over a long enough time period that the
vented material had collected air fluff as it dried.A small spot
of vented material had been deflected onto the LV driver control
board, bridging board vias and requiring scrubbing w alcohol to
remove.

One side of the inverter bridge was shorted through the switching
fets. When replaced, the recapped inverter ran normally under
lighter loads (below 500W).

I might mention that one of the unit\'s fans was dead, as well.

There is a small visible nonlinearity in the negative-going
sine output \'increasing\' edge, which looks increasingly like
undamped response as load increases. This (negative generating
pwm) was the side of the bridge that had failed previously.

At some point of increasing load, the negative going sine
output flat-tops - coincident with the positive going side
cutting out completely on its falling edge, then the unit
latches off to an overload limit.

I\'m not sure if this performance is aggravated by the loading
method - resistive through a variac, so any DC imbalance can
result in largish mag current imbalance.

Input to the low frequency AC limiting circuit seems normal,
though it is designed with a ridiculously high impedance -
a multimegohm RC filter from a 70milliohm sense resistor -
possibly to cut down on C size. It floats on the swinging
polarity of the negative bulk caps. Even some of my go-to
diff probes leak enough to make this a \'hands-off\' test point
while in operation, but what I can see of larger signal stuff
at 50:1 diff looks normal.

There are no schematics available for the 3x control cct boards,
which are basically 339s and 324s working with schmitt triggers
and flip flops. Gate drive is optoisolated x4 with DC restored
bootstrap power on each. LF AC reg (tl431)and fault signals are
optoisolated

A main board schematic is available on groups.io in Electronics101.
This shows that any control cctry components on the main board
are basically concerned with fan control or external comms.
Red detail indicates modifications not present on production
(this) units - again mainly fan control.

http://ve3ute.ca/query/Samlex_SSW-2000_main_board_schema.pdf

Any ideas or similar experience out there?

RL

Housekeeping supplies are reverse polarity protected, simple
and solid. AC voltage output LF amplitude is well regulated.

A solution is found accidentally - could get >1KW out of this
when a diff probe was attached between the AC vreg divider
point and the floating reference ground for that section.
That\'s between the junctions of R27/R28/R56 and the \'bottom\'
of VR1.

This is the rectified but unfiltered AC output, to provide feedback
for 60Hz voltage regulation.

Why? Who the hell knows. I tacked on a 100pf capacitor, burned the
thing in and booted it out the door.

RL
 
On Sun, 06 Nov 2022 11:16:09 -0500, legg <legg@nospam.magma.ca> wrote:

I\'ve got an SSW2000 repair that seems to be turning into a science
project.

The customer bought it cheap off Kijiji. It didn\'t \'go\'.

I found all 10 of the 30A automotive fuses blown, suggestting that
the last thing done to it was a reversed battery connection.
None of the DC inverter fets were damaged - This section proved to
run when a normal downsteam load was present.

The two main HV bulk caps were dried out and had vented oton the
wrapped heatsink. This over a long enough time period that the
vented material had collected air fluff as it dried.A small spot
of vented material had been deflected onto the LV driver control
board, bridging board vias and requiring scrubbing w alcohol to
remove.

One side of the inverter bridge was shorted through the switching
fets. When replaced, the recapped inverter ran normally under
lighter loads (below 500W).

I might mention that one of the unit\'s fans was dead, as well.

There is a small visible nonlinearity in the negative-going
sine output \'increasing\' edge, which looks increasingly like
undamped response as load increases. This (negative generating
pwm) was the side of the bridge that had failed previously.

At some point of increasing load, the negative going sine
output flat-tops - coincident with the positive going side
cutting out completely on its falling edge, then the unit
latches off to an overload limit.

I\'m not sure if this performance is aggravated by the loading
method - resistive through a variac, so any DC imbalance can
result in largish mag current imbalance.

Input to the low frequency AC limiting circuit seems normal,
though it is designed with a ridiculously high impedance -
a multimegohm RC filter from a 70milliohm sense resistor -
possibly to cut down on C size. It floats on the swinging
polarity of the negative bulk caps. Even some of my go-to
diff probes leak enough to make this a \'hands-off\' test point
while in operation, but what I can see of larger signal stuff
at 50:1 diff looks normal.

There are no schematics available for the 3x control cct boards,
which are basically 339s and 324s working with schmitt triggers
and flip flops. Gate drive is optoisolated x4 with DC restored
bootstrap power on each. LF AC reg (tl431)and fault signals are
optoisolated

A main board schematic is available on groups.io in Electronics101.
This shows that any control cctry components on the main board
are basically concerned with fan control or external comms.
Red detail indicates modifications not present on production
(this) units - again mainly fan control.

http://ve3ute.ca/query/Samlex_SSW-2000_main_board_schema.pdf

Any ideas or similar experience out there?

RL

Housekeeping supplies are reverse polarity protected, simple
and solid. AC voltage output LF amplitude is well regulated.

A solution is found accidentally - could get >1KW out of this
when a diff probe was attached between the AC vreg divider
point and the floating reference ground for that section.
That\'s between the junctions of R27/R28/R56 and the \'bottom\'
of VR1.

This is the rectified but unfiltered AC output, to provide feedback
for 60Hz voltage regulation.

Why? Who the hell knows. I tacked on a 100pf capacitor, burned the
thing in and booted it out the door.

RL
 

Welcome to EDABoard.com

Sponsor

Back
Top