Repair of Axpert Inverters : A Journey Started

[wael_fathe]

On 2023/11/05 at 1:33 AM, BritishRacingGreen said:

I tend to agree with you , there is closed loop control via the dsp but it is slower than a tight analogue close loop control. Further to this there is no dedicated current sensor on the battery side in order to rapidly detect short circuit currents. Wether such an addition may protect mosfets and igbts I dont really know.

detcting shorts in primary is  very fast...uc3842  based smps  and other  "current mode"  control   smps's   are all super

fast  when it  comes to protection   they die    due to  faulty  caps  more  than they  die   due to over  current  incidents 

[Coulomb]

6 hours ago, Mischel68 said:

TVS 1 to 12. It looks like a diode?

TVS stands for Transient Voltage Suppressor. These are like zener diodes, but optimised for handling high pulse power, usually 400 W or 600 W depending on the package size. They come in unidirectional and bidirectional; you probably want unidirectional (so install the right way around). Choose one with a standoff voltage rating about equal to the maximum expected voltage in normal (non transient) conditions.

[Mischel68]

4 hours ago, Coulomb said:

maximum expected voltage in normal (non transient) conditions.

Thanks for your answer.
Since the battery voltage is present at the MOSFETs, the value of the TVS diode should be around 60V. Or am I wrong in my thought.

[Coulomb]

4 hours ago, Mischel68 said:

Since the battery voltage is present at the MOSFETs, the value of the TVS diode should be around 60V.

Yes. The TVSs have various voltage ratings, the one you want is usually called the stand-off voltage.

[wael_fathe]

On 2023/11/04 at 11:00 AM, BritishRacingGreen said:

So to bring up the MAX system supplies without powering the battery terminals , I merely remove the SPS BAT+ cable , and feed into the main board with my current limited power supply. You can of course power the main board via feeding the SPS that is 'hot- wired with mains input . In fact I have done so originally , but I do prefer to use a current limited bench power supply instead.

 

the  units  i met  all  have   built in         " main smps   12 5 minus  12 "        and  bettery  sps 

but  i guess  feeding   dc  after  the " diode  or"     connections 

will prevent  fed back to battery what so ever

i think this can be done  to  sps configuration  where the "oring is  done  in the  battery side

and  can  be  also  done  in  sps  confirgurations  where the oring is done  after the  battery sps  

[wael_fathe]

On 2023/11/04 at 10:14 PM, razmasoom said:

i attached to secondary after diode

how   foolish  of me  ..yes  after diodes  and   no  voltage is going to back  fed

[Mischel68]

I asked the dealer in China again. The TVS diode is bipolar and 160V voltage.

[wael_fathe]

TODAY i got   good repair  story  that will  benifit  every one 

 

i received  an  inverter 48vdc 5500  watt  ,  the  symptoms is  burned  mosfets  and   dc-dc  igbts

i scaned  and  replaced  all  bad  fets  igbts  and  resistors   in both  primary and secondery

when i  turned the device on  ,,the inverter  on  but the  seems to be  restarting  continusly 

i  immediately  suspected  the main sps   when i checked  12 5  and -12  they all  ok  

i  checked the battery  sps  ,  its  output  is   fluctating    like  crazy ffrom 120 to 80vdc ...so big  ehhha  

i  brought my cigarette  and  coffee  as i know  i am  in the  right path 

logically what sps  is  feeding?  it  feeds  the main sps  and   the  soft start 

so   i  checked  the   mppt board  as the soft   start sps  existed there 

 

i find  a shorted  diode..,  this  certain diode used in mppt as main  boost diode  

and this  certain diode  i am sure i  cheked it   when i first opened   the inverter  it was  goood

 

i felt that its  replacment  wont bring any  fruit  ,  but  i did replace it any way

 

now   as  soon as  i turned on the  device  the diode  got shorted again  and not only   it shorted 

 

but   mppt  igbt  is  shorted as well  

 

....i know  i  was  close ....but   i should  avoid losing  new  diode and  new igbt  

 

i was not  sure what to do   ....i decided to   replace the diode  and  igbt  again  but   before doing so  i  replaced all

bus  500vdc  caps  .....when i  turned the  device on .. the diode and  igbt  not shorted  

 

and the  inverter  no longer  cycles   , it gives  me error 53..i  have  seen  this  error  before it is always  have  some thing to do  with  the  ac igbts ...so  i tested the   square  wave  signal  in all 4  igbts  and  find  2   missing  pulses   

i tracked  18vdc -5vdc  power  supplies  and  i  detcted   one  diode  is too hot  there   

 

i  tested  that diode i find it  shorted partly  ..i  turn  the device off and replaced the diode  ...

i powered  the  device  agian   and   mppt  igbt  and diode  blow  again 

 

so it seems  they  survive  when i replaced the caps  not because  i replace the caps  but 

  because    last  igbt  was fitted have taken  the  zener 18vdc  diode  with it ,

 

 prevnting 18vdc to reach to  the  copler that exist in mppt

 

that copler is  partly shorted   give him  18vdc  and he  will destroy ur igbts  ....a partly shorted copler will miss translate 18vdc into a  dc signal that is not on   nor  off   

 

 

i  quck   fix i  diconnected the  shown  cable  it is  responsible  for  sending the  18vdc to the mppt  board  

i removed bad  fet and bad diode from mppt board  

and  fixed new  zener 18vdc   diode  ...truned inverter on ..guess what.....the  device worked!!

 

-------------------------------------------------------

 

failure  analysis   ...the device  have a  dc-dc failure   and  have bad copler  

the bad  coplerand and  bad zener   loaded the sps  cuasing cpu to restart  ....although its not logical because main sps  seems ok 

with good  zener   the  18vdc travel to the mppt board  and  ruin the  igbts  there due to bad  copler ..

    with a partly shorted  zener  18vdc   voltage  is not  going to travel to mppt board   and the igbts will  survive  

 

but in other hand you wont have  your ac  as  missing 18vdc  is used by 

ac igbts  as  well

 

 

with picture  you will see the cable that you have to disconnct when repairing  i hope you love  my story  but have 2  questions 

 

i  did replace  bad t350  with good  t250 and it worked  

can i  replace  bad copler  for3150   with  t250?  will it work ?

the "suspected "  bad copler have  30k between its 18vdc pin and 5 vdc pin   

is it normal reading  ?   

the diode   in mppt  board  have the full  bus voltage  across it 360vdc is that normal?

Edited November 14, 2023 by wael_fathe

[AndrewF]

On 2022/11/09 at 3:34 PM, BritishRacingGreen said:

The PWN duty cycle is typically zero due to the fact that FB pin 2 sits at 5v reference voltage . If this pin is pulled low then the duty cycle will increase and when it reaches 0V , the duty cycle is at its maximum limit 50%. So the DSP can start the PSU by pulling FB low via opto-coupler interface U17.   The high voltage bus output will then produced.

So on my own journey with the Axpert 5kw , I have discovered that this reference voltage that   has mentioned here does not get pulled low by the DSP, is there another process that causes the DSP to pull pin 2 low on U17 thereby pulling FB pin 2 low on U16 . There is 12V present on pin 1 of U17 BTW, so pulling pin 2 low on U17 will do the trick...When the unit is witched on after some time the DC to DC  converter does work generating 425V , but the soft start does not precharge the BUS caps......The unit switches off after about 2 seconds when the 425V is generated on the BUS....I have removed the BUS caps for this part of the repair... 

 

Any thoughts please.....

[Coulomb]

11 hours ago, AndrewF said:

The unit switches off after about 2 seconds when the 425V is generated on the BUS....I have removed the BUS caps for this part of the repair...

I don't think it's a good idea to remove the bus capacitors. Without them, if and when the bus soft start power supply runs, the output will shoot quickly to a very high voltage, as that power supply has uncontrolled voltage and relies on the capacitors to ramp the voltage. The lack of bus capacitors may be complicating things.

If you manually pull pin 2 of the opto to digital earth, does that start the bus soft start power supply?

Perhaps check the ULN2003 (from lazy memory) on the control board, to see that it's passing through the instruction from the DSP to turn on the supply.

If it's OK, then it seems that the DSP isn't asking for the supply to come on, meaning it may be unhappy with something. I assume that you would be reporting any warnings or fault codes.

[wael_fathe]

can i  replace  bad copler  for3150   with  t250?  will it work ?

the "suspected "  bad copler have  30k between its 18vdc pin and 5 vdc pin   

is it normal reading  ?   @Coulomb

[razmasoom]

On 2023/09/06 at 10:36 PM, BritishRacingGreen said:

I have drilled down somewhat on this issue .  the 'spike' appears as a short spike , but I  stretched out the signal wider in time ,  the images below refers:

 

The Y scale is actually 10V per division

It is  not apparent in the image, but the bottom of the spike is actually clipped by the zeners. I have removed one zener , and the spike goes below -20V.

Note that this is with an IGBT fitted , far more damped . If I remove the igbt , the spike oscillation can clearly be seen, as below: please note this is without zener clamping:

 

and here is the same wave , but zener clamping restored. you can see the shaving at the bottom.

 

 

This to me is good news because one can observe the zener clamping on the scope without actually have to remove one to prove its ok.

So although the pulse reaches very close or on the +-20V absolute maximum rating of the igbt , the circuit under repair is correct , and I will abide by that. I have performed the same test on two different 5kW machines and they both expose the same waveforms.

 

EDIT : It is noteworthy to see that when the igbt is fitted , its capacitive load increases the switch off period from 48nS to 960ns.

EDIT2: This data is of course also useful  if we want to check whether  an  alternative / replacement IGBT is suitable in the machine.   

 

I think selection of replacement IGBT and MOSFET is only dependent on dead time between the pluse generated from sg3525 ic in output A and output B. I think all igbt/MOSFET suited for replacement if turnoff time is below the dead time of pulse (out A and out and its doesn't matter what's is it's capacitance characteristics. 

[Coulomb]

On 2023/11/19 at 5:00 PM, wael_fathe said:

can i  replace  bad copler  for3150   with  t250?  will it work ?

If you are asking can you replace a faulty HCPL-3150 opto-coupler with a HCPL-T250, I would say yes. The 3150 is a quite crude device, only 0.4 A drive, the T250 has 1.5 A drive. Other parameters seem similar.

On 2023/11/19 at 5:00 PM, wael_fathe said:

the "suspected "  bad copler have  30k between its 18vdc pin and 5 vdc pin

If you see anything less than infinite resistance between pins 2 or 3 (on the diode, input) side with any of pins 5-8 (on the gate driver side), then that's very bad, and the isolation has completely failed.

If you're saying you see this between pin 8 (Vcc on output) and pin 4 (Vee on output), then that's completely reasonable. I would build a quick test jig to find out if it's working or not. I would expect to see effectively a diode between Vcc and Vee, i.e. on a diode range with the multimeter positive probe on Vee and negative probe on Vcc, I'd expect to see about 0.35 - 0.6 V. In the other direction, I'd expect to see open circuit, even though the chip will draw some power when idle.

For the record, the ACPL-T350 is the most powerful device, rated at up to 2.5 A peak drive (2.0 A max recommended).

[Coulomb]

1 hour ago, razmasoom said:

I think all igbt/MOSFET suited for replacement if turnoff time is below the dead time of pulse (out A and out and its doesn't matter what's is it's capacitance characteristics.

I disagree. Even the mighty ACPL-T350 is only supposed to drive a maximum of 10 000 pF (10 nF) of gate capacitance. There are a few porky IGBTs that exceed this value, and would not work well in these inverters. See for example this post and the one following it.

Edited November 20, 2023 by Coulomb

[razmasoom]

6 hours ago, Coulomb said:

I disagree. Even the mighty ACPL-T350 is only supposed to drive a maximum of 10 000 pF (10 nF) of gate capacitance. There are a few porky IGBTs that exceed this value, and would not work well in these inverters. See for example this post and the one following it.

Sir but presently most of inverter do coming with STGW80H65dFB ( 10524pf)which has obviously higher capacitance then earlier used irgp4066d(4440pf).

[razmasoom]

Can anyone confirm the value of capacitor c4 and C110 out of the circuit. As mentioned in schematic it shows 10nf and both are in parallel. But when i made a test circuit for sg3525 and other the frequency not outputting around 38kHz rather it shows around 7kHZ . but when putting 2.2nf (as have in stock ) ie. total 4.4nf (as both are in parallel) almost reach around 28kHZ and are accordance 2F = 1/{0.7 * 3[R(d)] * R(t) * C(t)} .I have already assembled the unit of mine and is on running  condition , so cant test out of circuit. it should be near 1.7-1.8nf i think. 

Edited November 20, 2023 by razmasoom

[wael_fathe]

14 hours ago, Coulomb said:

14 hours ago, Coulomb said:

If you are asking can you replace a faulty HCPL-3150 opto-coupler with a HCPL-T250, I would say yes. The 3150 is a quite crude device, only 0.4 A drive, the T250 has 1.5 A drive. Other parameters seem similar.

If you see anything less than infinite resistance between pins 2 or 3 (on the diode, input) side with any of pins 5-8 (on the gate driver side), then that's very bad, and the isolation has completely failed.

If you're saying you see this between pin 8 (Vcc on output) and pin 4 (Vee on output), then that's completely reasonable. I would build a quick test jig to find out if it's working or not. I would expect to see effectively a diode between Vcc and Vee, i.e. on a diode range with the multimeter positive probe on Vee and negative probe on Vcc, I'd expect to see about 0.35 - 0.6 V. In the other direction, I'd expect to see open circuit, even though the chip will draw some power when idle.

For the record, the ACPL-T350 is the most powerful device, rated at up to 2.5 A peak drive (2.0 A max recommended).

thanks  for your   comprehensive  answer  ...alas  i ordered 100pces  3150  and 100 t350...if  i got this piece of  information before  i  would have made my order  all in  t350

 

[Coulomb]

23 hours ago, razmasoom said:

Sir but presently most of inverter do coming with STGW80H65dFB ( 10524pf)

Really? I was not aware.

Is that only clones or genuine Voltronics?

[razmasoom]

22 hours ago, Coulomb said:

Really? I was not aware.

Is that only clones or genuine Voltronics?

May be I own a clone one and its using stgw80h65dfb. But i have watched many videos and forum , they are using it . I think irgp4066 has been obsolete and no longer manufactured. Others may confirm usage of 80h65 in their inverter (genuine)

Edited November 22, 2023 by razmasoom

[AndrewF]

On 2022/11/09 at 3:34 PM, BritishRacingGreen said:

When I got to my senses I measured the bus voltage and it was sitting at 315V

According to  the soft start circuit should get to 315v, unless I totally misunderstand the context of this post,.

my soft start has already blown blown IGBTs , The voltage is exceeding 600V, is there any  modification that been done to the FB to regulate this voltage ...

[BritishRacingGreen]

3 hours ago, AndrewF said:

According to  the soft start circuit should get to 315v, unless I totally misunderstand the context of this post,.

my soft start has already blown blown IGBTs , The voltage is exceeding 600V, is there any  modification that been done to the FB to regulate this voltage ...

Hi Andrew , sorry to hear about your failure .

I have documented the manual control of bus soft start  previously as shown below :

 

I have highlighted the word of WARNING  in bold .

 

 

 

[trevorinrsa]

Folks

I have an Axpert 5KW Inverter that stopped working after the weekend storm in Joburg. Initially it had a D3 error (Battery overvolts) and I checked and the battery volts was 48 so not the issue there?

I have tried Nungu Solar who say they can repair on their web site but I cant get through on a phone!!

Can I ask you good folk for another reliable repairer in Gauteng for Axpert?

Thanks in advance?

Trevor

[Chris Louw]

 @trevorinrsa  Contact   BritishRacingGreen on this forum. He will help you if possible.

5 hours ago, trevorinrsa said:

Folks

I have an Axpert 5KW Inverter that stopped working after the weekend storm in Joburg. Initially it had a D3 error (Battery overvolts) and I checked and the battery volts was 48 so not the issue there?

I have tried Nungu Solar who say they can repair on their web site but I cant get through on a phone!!

Can I ask you good folk for another reliable repairer in Gauteng for Axpert?

Thanks in advance?

Trevor

 

[wael_fathe]

Hello ...what is the normal tempreture of output igbts heat sink 

I can sense it with my hand with 100 wattt load its a bit high tempreture ...since i cant tell how much the tempreture all i cab say it is bit higher than a warm heat sink...the output coil is however cold...strange since all ones i repaired coils is more heated than heat sink   can any one answer the inverter is 5k qmax clone probably

[Scorp007]

20 hours ago, wael_fathe said:

Hello ...what is the normal tempreture of output igbts heat sink 

I can sense it with my hand with 100 wattt load its a bit high tempreture ...since i cant tell how much the tempreture all i cab say it is bit higher than a warm heat sink...the output coil is however cold...strange since all ones i repaired coils is more heated than heat sink   can any one answer the inverter is 5k qmax clone probably

A normal temp is so hot that you cannot keep your finger there longer than 1sec. Not that it is just hot when you touch it.