Yaskawa VFD. transistor marking with L33 ??

[fastline]

Sort of a long shot here... I am trying to repair a Yaskawa VFD. The unit powers up normally but when E stop condition is is cleared, the VFD should fire a magnetic contactor that will allow 3 phase power to flow into the diode packs to charge the DC buss. This is not happening and after 4 sec, the VFD will show an error code that the DC buss has failed to charge properly. More like not at all. I took it all apart and confirmed the MC coil is fine and that coil is turned on by a small relay. That relay is turned on via 24VDC from the control board, or rather one side of the relay coil is held at 24VDC and the other side should pull low to ground to fire the relay coil. That does not happen either.

So that moved me to the actual control board in which it appears a proprietary microcontroller fires a 3 leg SMD transistor that has a marking "L33". It is VERY small but I believe the uC sends signal to the transistor and one side of it is held at ground potential and when fired, should connect ground at that transistor.

I will need to test a bit more it appears either the transistor is bad, or it is not receiving signal to turn on which could either be a bad microcontroller or maybe dirty power. It is very common in this older drives to have bad capacitors that cause issues. I can see a "few" were replaced but appears they went cheap and did not replace all of them, which there are only 8 on the entire control board.

The micro controller says Yaskawa on it and I have not been able to pull any data on it. Figures they had proprietary markings made up for them. I am temped to just replace all the caps first but would really like to source the problem before swapping random parts.

There is a small power supply for the board and nothing else shows errors. The PSU provides what appears to be 5V and 24V, in which the 24V is at 23.8V and 5V was at 5.10V ,there was another voltage at around 18V I believe. I did not test them for AC but I can. There is that possibility that the PSU has issues that are causing this. Something like this industrial VFD, caps become the very first concern.

But, assuming the transistor is of issue, is there any way to determine what the L33 is?

 

[Diver300]

My first approach would be to turn off, discharge the capacitors and see what the transistor junction voltages are. Many multimeters have a diode voltage function that helps with that.

From your description it seems that the transistor is either an NPN bipolar or an N-channel MOSFET. The emitter / source is likely to be connected to ground, and the collector / drain is likely to be connected to the relay coil. The base / gate will be driven by the microcontroller.

An NPN bipolar will have a junction voltage of around 0.6 V between base and emitter, and the same between the base and the collector. On a MOSFET, there will be a body diode so there will be an 0.6V drop between the source and drain. The readings can be complicated by other components, so disconnect the relay coil if you can while taking the readings.

The next thing to look at is the voltage on the base or gate of the transistor, when the relay is supposed to be on. If the is no voltage (and the transistor doesn't appear to be shorted) then the microcontroller isn't turning on the transistor, so the microcontroller is either faulty or, more likely, some external condition hasn't been met.

 

[NsrMagazin]

After identifying the transistor we will give you a replacement. The turn off part is very important, if you don't know how to measure while its on, you will risk yourself, the device or get wrong readings.

If you do decide to measure under voltage, you can check does the MCU output a voltage and is it "3.3VDC" or "5VDC". Also there should be a resistor connected from the MCU pin to ground(NMOS) or from the MCU pin to the base of the transistor(NPN BIPOLAR) if the explanation for how the device works is accurate.

 

[fastline]

https://www.weisd.com/store2/NTE2414.pdf

Here is what I believe to be the critter. I did diode checks in circuit and I tested two others that are identical. All test the same, but I cannot always buy readings while in circuit. There most certainly is 22K ohms between base and ground. Per the diagram, it appears the resistors act as the protective gate resistor in one shot to protect the uC ?

By taking it all apart, I was able to verify several things like the fact that the "field ground" is isolated by a capacitor so my powered up tests were not valid and I need to use a different ground. I am hoping to see the voltage output from the uC, which will at least make me feel better that it can be repaired. Otherwise if there is no output, my concern will be the uC channel is dead in which replacement is the only option though I am likely to consider a work around since all it has to do is trip that little transistor to make the whole thing run!

I need to double check the PSU that powers this board. I need to see what the voltages are and if there is appreciable ripple which could indicate PSU issues causing down the line problems. It does not appear the system has any error codes for the PSU.

I also directly fired the relay last night just to make sure it was in fact OK. the OEM spec is 8.3mA at 24VDC and I tested at 6.8mA at 20VDC. Works perfect!

To be right honest, I just cannot find much wrong here so I am REALLY hoping I find issues with the PSU today or that transistor is actually bad and I cannnot test it, because I don't want to think about dealing with a bad uC!!! That would just be my luck, but it seems reasonably protected. However, I think I would have selected an opto for power coupling for the relay as they seem superior for uC protection.

 

[kubeek]

It could also be a 3.3v regulator.

 

[NsrMagazin]

I have the feeling something is missing.
Can you please post a picture of the transistor and MCU? The more details the better.

 

[rjenkinsgb]

It sounds like you are describing a "soft start" function.

Most inverter systems rated above trivial power levels have a two-stage startup; the first stage has resistors connected in line with the power to the main rectifiers, which allow the reservoir caps to charge at a moderate current rather than there being a high-current surge.

Once the caps are charged to near their full voltage a bypass contactor shorts out the limiting resistors so the device can work at it's full rated power.


The implication from the "bus not charging" message is that there is a fault in the power stage, so current is being drawn before the bypass contactor goes in - and the reservoir caps are being held at a low voltage.


Start with the output power device / devices and check the DC feed to that for shorts, eithert direct or to any output terminals.


(Also locate and check the charging resistors - if they have failed it would give a similar error).

 

[fastline]

Missing? like numbers or components? I feel I traced the PCB reliably. Also, the transistor is so small that the only way you can see anything is with a microscope and I don't have a way to take a picture with mine. Also, the uC is a 160pin chip that says "Yaskawa" on it. I tried to chase the numbers on it but I literally get 3 hits on google. Yaskawa is not in the chip business so they must have someone relabeling components for them.

 

[fastline]

rjenkinsgb
The way this one works is when there is an Estop condition, the DC buss powers down. That is common even on Fanuc drives. The only connection of 3ph power to the DC bank is through a "line reactor" they call it. It looks almost like a 3ph transformer but I believe it to be a choke that will limit inrush. It is a pretty substantial device!

The system should at least try by closing the MC contactor. I am pretty sure the system attempts to because it waits 4 sec after resetting before it throws the "fail to charge" alarm.

 

[rjenkinsgb]

I've just checked a Yaskawa VFD technical manual.
They definitely have a resistor plus contactor setup; it's mentioned in the diagnostics for an undervoltage fault.

The delay is the time the controller allows for the caps to charge through the limiting resistors.

Some drives have three resistors before the rectifier to distribute the power dissipation, some have a single resistor and contact after the rectifier.

One example, re. VFDs in general:
http://www.gozuk.com/blog/vfd-basics-864930.html

 

[throbscottle]

The 2007 Turuta SMD codebook gives L33 as being either FA1A4M or GA1A4M
They are both a NPN "digital" transistor containing 2 x 10k bias resistors, 60v, 100mA, 200mW.
FA1A4M has SC59 case,
GA1A4M has SC70 case but I think you can't tell the difference by looking.

 

[fastline]

The big wires go from MC, to line reactor, then to the power diodes. That is how it is wired. What am I missing here? If the reactor is not shown in the diagram, I can post a pic. There is no way you can call it a resistor.

 

[rjenkinsgb]

Do the power diodes connect directly to both positive and negative of the main caps?

Are there any other smaller connection on the contactor main terminals or power diode AC side?

What model number is the VFD? I may be able to find further info.
Photos of the all the power circuits from inputs to capacitor bank would be very helpful as well.

If is has an internal main power contactor, there may be a second small contactor or relay involved in the soft start charging circuit. There are hundreds of variations of arrangements between makes and models of drives.

 

[fastline]

Model is 626VM3. pg 39 is the schematic, page 58 though explains the Estop condition VERY clearly and I feel I am on point with this. There is only one BIG MC that sends power to the DC buss diodes and those diode packs then connect directly to a main buss bar that feeds the IGBTs.

If you read on page 58, it states that the drive will not be ready for 2.5sec which is the time to charge the buss. On mine, it errors out at 4sec because it knows there is no voltage on the DC buss because the MC never closes.

As to the explanation for the error code talking about a "resistor", it ain't there! The only thing you have is bleed resistors to ensure the DC buss discharges. The line reactor is responsible for current limiting to the cap bank and diodes and I know that has nothing to do with it because the MC won't even close.

In all actuality, if I just hook a 24V source to the relay for testing, I bet the drive would work! But, I will comb the circuit more to determine if the uC output works or not. It is possible the system is detecting something strange or maybe it is getting dirty power?

 

[fastline]

I have reconnected all the hardware in the drive and retested. There is certainly no output from the uC. There is another identical transistor and connections next to this one and when "reset" is toggled which should initiate the drive and charge the DC buss, it receives a 5VDC signal and the feedback of 24VDC on it pulls down to 1V indicating it functions as intended.

The problematic device has absolutely zero output, not even a whisper. I have tested that leg to ground to ensure it is not shorted somehow and cannot find much. When power is removed, the leg from the uC is at 2.4Mohms. When powered, it reads 25ohms, which I felt was a bit odd.


I feel I have reliably tracked the trace in the PCB and it goes directly from the output of the uC to the base of that transistor and that is it.

I also tested what I believe to be power to the uC and it sits at 5.09VDC with 1mV of AC ripple. I also tested the power supply for this board and voltages were 5.10, 14.98, and 23.84, all with 1-2mV of ripple. PSU seems to be just fine.

I would say if Yaskawa tells me to replace the drive, I am going to hot wire it and maybe use the source power from the neighboring transistor to drive this one and cut the trace going back to the uC. I don't like doing hacks like that but I cannot get a PCB diagram and pretty certain Yaskawa will hold out on the needed info to figure this out. For all I know, there is a spare output on the uC that could be used and I would bet they planned for that.

 

[rjenkinsgb]

See page 211 of the manual, the faultfinding section.
https://www.yaskawa.com/delegate/ge...6C&cmd=documents&documentName=SIE-S626-6C.pdf

The second cause is "Failure of the charge current suppression resistor"
The main caps are supposed to charge within the time before the main contactor pull in.


If you manually override that with the caps not charged, you are likely to destroy the rectifier stage due to extreme inrush currents.

There are warnings in the manual that repeated starts will age the charging resistor(s); they are over-run on the basis they can normally cool between starts. On older Yaskawa drives they are listed as a service part but it looks like they now see that as an excuse to charge for a repair...


You mention another relay driver circuit is switching on when you try to start the drive. That will be the initial power on that should start the charge sequence.

Trace the relay contact circuit for the one that is not pulling in - it's either across the charging resistor or controls a larger contactor/relay that has contacts across the charge resistor.

Failing that, look what the relay that is going in applies power to - there may be a small auxiliary rectifier that feeds the charging resistor; that's sometimes used with regenerative PSUs as the main "rectifier" is actually a bank of transistors.


This is Yaskawas resistor notes for one series of drives (I was looking for the spare part for yours but have not found it yet):
https://www.yaskawa.com/delegate/ge...c1-2adf-4737-add7-69542dd0b463/DRV-5WHLKD.pdf


Another manual for a different series, which gives power rating, resistance and part numbers for various "soft charge" resistors for different sizes of drive; see page 54:
https://www.yaskawa.com/delegate/ge...ents&documentName=PP.P5G5.03.Troubleshoot.pdf

It's got one in there somewhere!

edit: The output resistance readings on the MCU are normal. The output stage FETs are off when there is no power but the output is being actively driven low when powered up, so giving a low reading.

 

[fastline]

I'd like to see where in the schematic it shows the resistors? How are they powered? They are not even on the schematic are they?

Also, what is the purpose of the reactor in there? That is shown on the schematic as ACL and it is certainly there.



At least for the moment, I have had to direct effort to fixing the now dead PSU in it. Made a new thread on that one.

 

[rjenkinsgb]

The "schematics" in the manual are only basic block diagrams, not detailed drawings.
Over the years there has been a steady reduction in technical info in product manuals, from often having full detail drawings of every part or circuits board in the 60s / 70s to minimal block level stuff now - the makers want to sell more, rather than producing a product that can be maintained forever...

I'd guess the resistors are connected to the block labelled "Power supply detection" or the next one that controls MC.


The reactor is likely for both commutation and suppression.
It's a regen drive - when the motor is being braked it puts energy back in to the DC bus and rather than being dumped in to a brake resistor and wasted, it's fed synchronously back to the mains supply.

That will produce noise and probably current transients (at the microsecond scale) as the current is switched between devices, which the inductor blocks from the external supply.

 

[fastline]

I am going to dig harder for that charge. Would be interested in your input on the PSU issue.

I do know the main PC board has a single mechanical relay on it that switches when you hit reset to initiate the drive. I might have to look closer at that. Even an official Yaskawa video mentions a special charge circuit before the mains power is applied. It has to be there!

 

[fastline]

Well..... I found the elusive soft charge circuit! Surprisingly, even when looking VERY hard for it, it still took 30min to find. The wires come off the back side of the main contactor and hiding in the back side is 2, 10ohm resistors and a smaller mag contactor. I chased the coil wires and found that the precise other L33 transistor that I tested and confirmed to work correctly is in fact the one that drives that soft charge coil. I am totally baffled at the moment because not only was I able to see 24V at that transistor as feedback from the mag coil, I confirmed that it pulled low which means it should have fired. I am going to have to see if the 24V source may be at issue here and possibly that PSU I am working on. Just not sure. I do know all the main components are fine and there is some small gremlin trying to test my sanity.