1 hour ago, wael_fathe said:

i think   the  forced  method needs the entire  control board  to be  removed  but i have  seen units   with  un  removable  control  board   and i have  seen   units   with soldered   cpu  ?

Ah, yes, for those, you would need to be more creative. I'd still use the forced method, but I'd disable the 5 V power supply, either by removing D69 or just powering up the +12V and -12V with a dual power supply. If there is no 5V, then the processor won't have power, so no outputs will get pulled down. Without the main power supply running, TX7 won't get powered, so you'll have to test the IGBT gate drives separately with a third bench power supply. 

16 hours ago, Coulomb said:

Ah, yes, for those, you would need to be more creative. I'd still use the forced method, but I'd disable the 5 V power supply, either by removing D69 or just powering up the +12V and -12V with a dual power supply. If there is no 5V, then the processor won't have power, so no outputs will get pulled down. Without the main power supply running, TX7 won't get powered, so you'll have to test the IGBT gate drives separately with a third bench power supply. 

ok  let  me  make  sort of  general  steps  for doing  so 

 

1-  remove the   cpu  board  or  remove    the 5vdc  that comes from the standby power supply or  the so called main sps 

2-  power  up the   device  using on off switch this  switch will  switch on the   main sps  

3-   short  the coupler  near by  the sg3525  or  pull its  sd  pin  to what ever level  guanrnatee  the work for sg3525

5-  test primary  fet  from  gate to source   looking for   square  wave  in the range of  kh  (all fets  removes  and unsoldered )

6-  check  igbt  driver  from gate to source   look for  square  wave the  range of  kh  (all  igbt  rectifiers  have to be removed   unsoldered )

----------------------------hope it right will see your feedback on this  mr  coulomb-

 

 

Edited February 19, 20233 yr by wael_fathe

3 hours ago, wael_fathe said:

(all  igbt  rectifiers  have to be removed   unsoldered )

Test that there is nothing on the DC bus; if so, there is no need to remove the IGBTs.

If you have enough bench power supplies, you can power the ±12 V supplies and one IGBT supply and not have anything on the battery bus; in that case, you don't have to remove the MOSFETs either.

Edit: otherwise, all else is good.

Edited February 19, 20233 yr by Coulomb

so  with  dual  supply    12-+ 

and  with  cpu board  diabled or removed   and with out bother turn  the power  bottom on  

and with out  bother   remove  fets out or  remove  igbts out   

we can go on short method and test for  drive in both primary  fet and secondery  igbt(rectifiers)  ??

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

but  with the  12vdc  or24(according to  inverter's dc input )   plugged in  

fear factor is in !   we  must make  sure  cpu  is  disabled

all pri   fet removed all rect igbt  removed   caps  at  +hv  bus  have to be  checked  for  high voltages 

and  only then       ....igbt  and fet can  be  tested for square  wave?

 

 

2  diffrent  routes for the detective  to find the cuplrit  wit  each one offering   diffrent  degrees of            easiness  vs   riskness 

Edited February 20, 20233 yr by wael_fathe

On 2023/01/30 at 8:21 AM, Coulomb said:

I would say so. The only problem would be if the gate drivers can't handle the extra gate capacitance of the larger GP4066D part (140 A versus 100 A for the GP4063D part). [ Edit: And if the power supply can handle the slight extra drive current; I imagine it would. ]

Does the 3000W model use the same gate drivers (ACPL-T350, marked T350) as the 5 kVA models? If so, you should be fine using the stronger GP4066D chips.

And you can take that poor GP4063D off the chopping block 😉🪓

 

 

in the inverters  that use mosfet   output,  i have  replaced  many times  

47n60  with   irf460     and   vice  versa    with out any heat issues   . 

if mr  combos  agree   mosfet  is more  friendly  when it comes to replacement  than  igbt

3 hours ago, wael_fathe said:

so  with  dual  supply    12-+ 

and  with  cpu board  diabled or removed   and with out bother turn  the power  bottom on  

and with out  bother   remove  fets out or  remove  igbts out   

we can go on short method and test for  drive in both primary  fet and secondery  igbt(rectifiers)  ??

You need to make sure that you don't connect AC-in, PV-in, or battery as well.

Yes, you should be able to test the MOSFET drive that way.

And, my apologies, you should be able to check the DC-DC IGBT gate drive that way too, without needing a third power supply. This is because the DC-DC IGBT gate drivers all operate off ±12 V.

But since TX7 will not be active (no main power supply), in order to test the full bridge (main DC-AC converter) gate drives, you will need a third bench power supply to replace each of the three outputs of TX7 to test the IGBT gate drive. You also need a way of pulling each IGBT gate drive signal (e.g. CN11 pin 2, pin 3...) and note that the gate drive goes from about -5 V to about +13 V when you pull that input low, and back again when you let it float high.

4 hours ago, wael_fathe said:

but  with the  12vdc  or24(according to  inverter's dc input )   plugged in  

fear factor is in !   we  must make  sure  cpu  is  disabled...

If you're referring to the battery terminals (which could be 12 V, 24 V, or 48 V), then yes, you need all those protections. But there is no need to connect anything to the battery terminals for this testing.

 a Word of Warning when testing the Bus Soft Start Circuit Manually

In previous posts I have suggested that one can debug a faulty bus soft start circuit by removing the control board and exciting the soft start circuit manually .  However one must keep in mind that when the circuit suddenly start to work , the bus voltage will reach some uncomfortable high voltages. The reason is the control board knows where to stop the soft start , but when we are enabling the soft start manually , there no stopping . I have fortunately had a meter on the bus , and when it reached close to 500V ! , I switched off.  So one way of ensuring safety is to remove the diode in the high voltage transformer secondary , then debug your faulty circuit in comfort.

 

Max 7.2 Schematics

So far I have found a number of errors and missing labelling etc in my schematics , apart from what Shadders and Coulomb have already pointed out . So its basically time to revisit them , and also publish a sort of prelim version  in original Kicad source format  for anyone that would like to add or do modifications . Of course I would like to in return benefit from such changes etc as well.

 

Hi @Coulomb I see that older Axpert 5kw models (MKS2) has R90 and D32 fitted. Yet on newer models (MKS4) they are omitted and R91 is 47R instead of 22R . Has this got to do with the type of IGBT fitted. Perhaps on the older models the IGBT input capacitance is higher than newer types .  My main question is : do think its ok to insert an MKS4 IGBT on the MKS2 board and keep D32,R90 and R91 intact as it is , or should I match the MKS4 drive components?

5 hours ago, BritishRacingGreen said:

Yet on newer models (MKS4) they are omitted and R91 is 47R instead of 22R . Has this got to do with the type of IGBT fitted.

This is on the edge of my expertise, but yes, I believe so.

5 hours ago, BritishRacingGreen said:

My main question is : do think its ok to insert an MKS4 IGBT on the MKS2 board and keep D32,R90 and R91 intact as it is , or should I match the MKS4 drive components?

I believe that you should match drive components, i.e. in this case change R91 to 47Ω and omit one or both of D32 or R90. But as I say, I'm no expert with IGBT gate circuits.

If you're referring to the battery terminals (which could be 12 V, 24 V, or 48 V), then yes, you need all those protections. But there is no need to connect anything to the battery terminals for this testing.EER

 

 

I dont get this

i have  2  methods 

 

one with  connecting the battery

one   without  connecting the battery

 

the     " one without"   ,   we  use instead  a  dual power supply 12+-

but you told me  for  the second method also dont  connect any thing to battery   terminals!

 

 

the  second  method invlolves  unsolder fets  and rect  igbts 

connect  supply dc input 24vdc 

diable  cpu    ,drag  sd pin  3525  to logic low  , power on the main sps  by turning on  the  power  buttom  

scope  for  wave forms !    so it does  involve connecting the power supply   !

 

thanks 

 

10 hours ago, wael_fathe said:

the     " one without"   ,   we  use instead  a  dual power supply 12+-

but you told me  for  the second method also dont  connect any thing to battery   terminals!

The ±12 V supply is totally separate and isolated from the battery terminals. Powering the +12 V and the -12 V supplies (two separate supplies, both referenced to "ground", so  24 V total) does not power the battery terminals, hence even if some MOSFETs are shorted or not driven correctly, it won't explode anything or produce dangerous voltages.

hello people 

first i would like to give  special thanks  to mr  colombos and  all  people in this  great  forum for thier support  and  togetherness 

 

today i carry to you all  guy success story  so please listen 

 

qmax 24vdc 3000 watt come in for repair   the  problem was 09 error

i opened the inverter  and  find the  previous  techncian  have  replaced all   h-bridge  and  rectifer igbts by  bad  quality igbts

the primary fets  and thier  drive  circuits cheked ok 

and i also find  few burned  zeners and  reisstors  in recctifier  igbts  gate  cicruits 

 

i assume  that  the  inverter   first come in  with onlyhbridge   igbts problem 

after technician  replace  all  igbts  with bad qualty one ,

he   introduced  another  failure   ,   since all  rectfier igbts  went bad

 

 

i started  by cleaning and replacing eevery  bad  zener  diode  reisstor  in rectifier  igbt   gate  circuits   ,

 

after i  finished i checked the  h bridge igbts  they seems all fine

 

i installed  new igbts  for  the  rectifier  igbts  and  fired the  units on 

 

error  09  appaer again and  one  igbt in the  h-bridge   shorted 

 

 

i felts that i  was  too  hasty  ,  i should  have checked the  busvoltage    dc   and  the pulse  drive  for the  hbridge igbts    before i turn  the device on 

 

 

so i desoldered  the  h-bridge  igbts  trun on device  and tested  the bus,   it was normal  at 370vdc  also  the   drive  gate   was ok   square  wave  , it stays for  about  3  seconds then the  device shutdown and report 53  error    

but i noticed that one drive is missing    i  shitdown   device and trace  back to find  47 ohm resistor  opened

 

i  solder new 47 ohm  in   fired the unit   again checked the  wave forms   they   all  check ok  ,   i have  feeling that with new  igbts  it should  work  fine 

 

 

 

true  enough  i installed  k50h60   and turn the  device on  it all worked  fine    i even lload  the   qmax   3k  inverter  with 60w  lamps   it  glowed   nicely 

 

i  cant  believe i did it    the gates  of the rectifer  igbts   have lots of  bad  reistors  and  zener     i did it i did it i am  happy   !

about 16  picture is loaded  for the repair  story you can easily  download the picture  to see    how  i did it 

each picture  has  number  follow the numbers from 1  to 16  to see how  i progressivly   repair the inverter 

.https://www.mediafire.com/file/axldufan39ufoqg/repair+story.rar/file

 

Edited March 4, 20233 yr by wael_fathe

6 hours ago, wael_fathe said:

hello people 

first i would like to give  special thanks  to mr  colombos and  all  people in this  great  forum for thier support  and  togetherness 

 

today i carry to you all  guy success story  so please listen 

 

qmax 24vdc 3000 watt come in for repair   the  problem was 09 error

i opened the inverter  and  find the  previous  techncian  have  replaced all   h-bridge  and  rectifer igbts by  bad  quality igbts

the primary fets  and thier  drive  circuits cheked ok 

and i also find  few burned  zeners and  reisstors  in recctifier  igbts  gate  cicruits 

 

i assume  that  the  inverter   first come in  with onlyhbridge   igbts problem 

after technician  replace  all  igbts  with bad qualty one ,

he   introduced  another  failure   ,   since all  rectfier igbts  went bad

 

 

i started  by cleaning and replacing eevery  bad  zener  diode  reisstor  in rectifier  igbt   gate  circuits   ,

 

after i  finished i checked the  h bridge igbts  they seems all fine

 

i installed  new igbts  for  the  rectifier  igbts  and  fired the  units on 

 

error  09  appaer again and  one  igbt in the  h-bridge   shorted 

 

 

i felts that i  was  too  hasty  ,  i should  have checked the  busvoltage    dc   and  the pulse  drive  for the  hbridge igbts    before i turn  the device on 

 

 

so i desoldered  the  h-bridge  igbts  trun on device  and tested  the bus,   it was normal  at 370vdc  also  the   drive  gate   was ok   square  wave  , it stays for  about  3  seconds then the  device shutdown and report 53  error    

but i noticed that one drive is missing    i  shitdown   device and trace  back to find  47 ohm resistor  opened

 

i  solder new 47 ohm  in   fired the unit   again checked the  wave forms   they   all  check ok  ,   i have  feeling that with new  igbts  it should  work  fine 

 

 

 

true  enough  i installed  k50h60   and turn the  device on  it all worked  fine    i even lload  the   qmax   3k  inverter  with 60w  lamps   it  glowed   nicely 

 

i  cant  believe i did it    the gates  of the rectifer  igbts   have lots of  bad  reistors  and  zener     i did it i did it i am  happy   !

about 16  picture is loaded  for the repair  story you can easily  download the picture  to see    how  i did it 

each picture  has  number  follow the numbers from 1  to 16  to see how  i progressivly   repair the inverter 

.https://www.mediafire.com/file/axldufan39ufoqg/repair+story.rar/file

 

Great, I love it when a plan comes together. 👌

Well done. Do you check your IGBT gate drive signals with oscilloscope? 

Edited March 5, 20233 yr by BritishRacingGreen

Hi @Coulomb just some small notes.

On the 5kw main boards, i have noticed that Voltronics use Jianghai CD297 spec 470u500v  caps on HV dc bus thru out older boards and brand new ones. The spec is actually impressive : 7000hrs@105.  

The battery bus caps has undergone changes over the years. Older boards have  4 x 3300u63v  thru out. However newer boards have 2 x 2800u80v  plus 2 x 3300u80v.  80v is a great upgrade, 63v too close to 56V operating. But interesting, why have slightly lower capacitance at the start of bus, thsn at the end of the bus.?  Or was it merely driven by procurement challenges? 

 

 

 

 

Edited March 5, 20233 yr by BritishRacingGreen

21 minutes ago, BritishRacingGreen said:

why have slightly lower capacitance at the start of bus, thsn at the end of the bus.? 

I've noticed that too. It seems to happen too often to be procurement, though that's still possible.

My suspicion is that the lower capacitance part has a lower ESR and/or higher ripple current rating than the other part. The actual capacitance is more or less meaningless in this application; 100 μF would possibly suffice IF you could get a low enough ESR and high enough ripple current rating. So the capacitance is mostly a side effect of getting enough surface area to spread around the pulse loads, I think.

6 hours ago, BritishRacingGreen said:

Great, I love it when a plan comes together. 👌

Well done. Do you check your IGBT gate drive signals with oscilloscope? 

please  download  rar it contains   alot of   repair pictures  

this is  scope  picture  go on and  download  all  repair  files from  here you will enjoy  them  they  are  numbered  1.2.3.4  etc etc    they have  comments on them  you can follow  the entire repair story  just  by  folliwing them 

 

 https://www.mediafire.com/file/axldufan39ufoqg/repair+story.rar/file

 

Edited March 5, 20233 yr by wael_fathe

Hi @Coulomb I have some difficulty with Axpert MKS4  main boards where there is a loss of earth bonding to output neutral . The image below illustrates the connection of Earth to the normally closed contact of RLY2.  You will see the big copper pour representing earth in the bottom right hand corner of the image. As you can see there is only a thin PCB track from this earth to the RLY . This track has incidentally burned and discontinued as you will notice. I attribute this to some earth fault currents  from the load , but the thin track cannot cope with such currents.

 

 

In contrast ,if you look at the MKS3 main board , you will notice that the PCB layout is nearly identical , except the earth track to RLY2 N/C contact . This is a proper thick copper pour . Incidentally the track on MKS4 represent the outline of this pour . So I wonder if this is a PCB layout error , and if so , its a nasty one.

 

Edited March 7, 20233 yr by BritishRacingGreen

10 hours ago, BritishRacingGreen said:

Hi @Coulomb I have some difficulty with Axpert MKS4  main boards where there is a loss of earth bonding to output neutral . The image below illustrates the connection of Earth to the normally closed contact of RLY2.  You will see the big copper pour representing earth in the bottom left hand corner of the image. As you can see there is only a thin PCB track from this earth to the RLY . This track has incidentally burned and discontinued as you will notice. I attribute this to some earth fault currents  from the load , but the thin track cannot cope with such currents.

 

 

In contrast ,if you look at the MKS3 main board , you will notice that the PCB layout is nearly identical , except the earth track to RLY2 N/C contact . This is a proper thick copper pour . Incidentally the track on MKS4 represent the outline of this pour . So I wonder if this is a PCB layout error , and if so , its a nasty one.

 

I have also noticed that the older MKS2 model main boards does not provide any earth bonding.

22 hours ago, BritishRacingGreen said:

So I wonder if this is a PCB layout error , and if so , its a nasty one.

It sure does. So it's not a trick of the light or anything, the trace really is that thin? There is no copper in that dark green area at all?

The track that's left (presumably after some sort of over-current event) looks thinner than even the thinnest "data" track.

Ick. Well spotted.

11 hours ago, BritishRacingGreen said:

I have also noticed that the older MKS2 model main boards does not provide any earth bonding.

Originally, Voltronic were oblivious to the need (as some other inverter manufacturers appear to be). So they didn't have the extra relay contact until someone from the AEVA forum contacted them and told them that they could have a safety issue. Every model from that time on on came with the extra contact.

23 hours ago, Coulomb said:

So it's not a trick of the light or anything, the trace really is that thin? There is no copper in that dark green area at all?

I removed RLY2 for this purpose :

As you can see the track routed from earth to resting contact terminal  3  is paper thin. On  MKS3 this is a wide copper pour. 

Edited March 8, 20233 yr by BritishRacingGreen

On 2023/03/07 at 6:04 PM, Coulomb said:

So they didn't have the extra relay contact until someone from the AEVA forum contacted them and told them that they could have a safety issue. Every model from that time on on came with the extra contact.

Thats thats super, we are now benefitting from except the MKS4 which has dodgy connection. 

Edited March 8, 20233 yr by BritishRacingGreen

22 minutes ago, BritishRacingGreen said:

I removed RLY2 for this purpose :

As you can see the track routed from earth to resting contact terminal  3  is paper thin. On  MKS3 this is a wide copper pour. 

When that track blows away, the fault is sadly undetectable. Not a big issue in other countries,but in South Africa with up to 8 hrs loadshedding a day, the probability of  someone not being protected by ELD is so much higher. Furthermore one cannot  ask owner to perform evaluation test of ELD via test button regularly , because test function that trips the ELD only works on current difference, and does not prove that the residual current is actually an earth current. 

EDIT: I actually instructed a owner to introduce a hard bond on the load, after we detected non conformance of MKS4.  Fortunately he has no upstream ELD on the grid connection. 

This is strictly against the rules as we have a bond on one than more location, but owners safety is more of my concern than this violation. 

Edited March 8, 20233 yr by BritishRacingGreen

@Coulomb some months ago you made an interesting comment regarding the MAX7. 2. Your view was that the 7.2 was the learner, or a shoe-in design for the 10kw MAX that  Voltronic had actually in mind.

I have evidence now to believe that you have got to be correct. I have a 10KW  OG10(MAX) main board with me and on face value the only difference is uprated bus and battery capacitors (MPPT left aside, dunno if there is difference) 

So if you look at the MAX10, then the drive train is double that of the 5kw machine. The IGBTs are dual, the HF transformer also dual etc. 

But the 7.2 matches the OG10 exactly except caps. 

Now 7.2 adds only 2.2 kw extra to the 5, yet it is doubled upgraded. 

 

This is why the 7.2 is such  robust, its technically overkill. I just hope Voltronics does not discontinue it. 

EDIT: this all does not infer that the OG10 is not capable to deliver 10kw reliably,  it most probably is. Time will tell. 

 

 

Edited March 15, 20233 yr by BritishRacingGreen