Blown X1 Mill speed control

[Oliver LindleyParticipant @oliverlindley63277]

[Oliver LindleyParticipant @oliverlindley63277]

I have a clarke mini mill (seig X1 variant), and the speed control has failed.  Would someone who has one be able to identify the main transistor for me please? (it’s the 3 legged part mounted to a heatsink on the speed controller board).  When it blew the part was damaged so that it was not possible to read the part number.  I replaced the part with an IGBT that i had knocking around, and it worked OK, but the speed fluctuted under load until it finally failed.  Obviously i would like to get the thing working properly again, so any help would be appreciated

[Peter G. ShawParticipant @peterg-shaw75338]

Hi,

 

Somewhere I read that Machine Mart, who supply a lot of Clarke stuff, have a very good spares parts catalogue. Perhaps that could be a good place to try.

 

A few thoughts are that the maximum working voltage of the device isn’t high enough. Again I believe I read something about this in conjunction with Arc Euro Trade and the X2 series mills. Perhaps there’s another avenue to explore.

 

Also, what about the power handling capacity?

 

Regards,

 

Peter

[Oliver LindleyParticipant @oliverlindley63277]

Hi Peter

 

Thanks for the advice, unfortunately Machine mart can only supply a complete replacement speed controller. At around £80 this is significantly more than the cost of the blown part – should be around £5-£10 – so I would rather try and fix it first.

 

The IGBT that i put in as a replacement was a 600V part, good for 25A, so i’m unsure why it went, but without knowing the topology of the speed controller i can’t make an informed judgement.  I plan to spend some time reverse engineering the board so that I can understand more about the design – like everything else on the mill it has been made to a price, and there is plenty of scope for improvement. 

 

[Keith PetleyParticipant @keithpetley53472]

Hi Oliver,

   The Little Machine Shop sells STW20NK50Z (IRFP460) mosfets for this purpose. Maybe you can find them locally.

 

http://www.littlemachineshop.com/products/product_view.php?ProductID=1221&category=

 

Regards,

 

Keith 

 

[Steve GarnettParticipant @stevegarnett62550]

You can get IRFP460’s from Farnell – don’t cost very much at all. Although I just checked the handling charge, and that’s the most offputting thing about dealing with them and small quantities. What worries me most about this though is that you’ve reported a blown device, the machine doesn’t run smoothly, so I think that probably, you should find out what’s really behind all of this – because at the moment it sounds as though the blown driver device is a symptom, not a cause…

Edited By Steve Garnett on 07/02/2010 23:19:30

[John BaguleyParticipant @johnbaguley78655]

Hi Oliver,

 

last year I repaired a couple of driver boards for the Machine Mart mini lathe (C3 etc) for a friend at the club. He was on his third board and getting fed up with the cost of the replacement from MM. I replaced the output mosfets on both with a higher spec version from RS and so far so good. The originals were the standard IRFP460 but I replaced them with IRFP22N50A which have a higher rating. The probem with the original IRFP460 is that they can’t stand the current drawn if the motor stalls under load. By the time the fuse blows, it’s too late!

 

John

[Steve GarnettParticipant @stevegarnett62550]

Dunno why anybody bothers with fuses and MOSFETS…

 

If an IRFP22N50A doesn’t blow up, well that’s fine – but it strikes me that the spec isn’t that much higher than for the IRFP460 really, so is the whole thing running just on the edge when it comes to stall current? On the face of it, it sounds as though some more thought ought to be given to the power supply to limit the stall current to a safe level.

 

[Oliver LindleyParticipant @oliverlindley63277]

Thanks to everyone for your help – I’ll get some MOSFETs.

 

Steve – I think the failure is something inherent in the design – looking on google, there are many people around the world with the same problem.  I don’t think it is the stall current that kills it – motor DC resistance is 24ohms, at 180V that’s 7.5A.  The MOSFETs are rated at 20A.  Much more likely is that the back EMF from a sudden stall causes the MOSFET to avalanche and without very careful design this can lead to failure of the device – not instantly, but each time the body diode breaks down, it is weakened until it eventually fails short.  The electrons literally punch through the oxide layer.

 

I have found a few parts with much higher dv/dt and repetitive avalanche ratings, so i may give one of those a try, and fall back on the IRFP22N50A if I am wrong. 

[Steve GarnettParticipant @stevegarnett62550]

Posted by Oliver Lindley on 08/02/2010 22:33:01: 

Steve – I think the failure is something inherent in the design – looking on google, there are many people around the world with the same problem.  I don’t think it is the stall current that kills it – motor DC resistance is 24ohms, at 180V that’s 7.5A.  The MOSFETs are rated at 20A.  Much more likely is that the back EMF from a sudden stall causes the MOSFET to avalanche and without very careful design this can lead to failure of the device – not instantly, but each time the body diode breaks down, it is weakened until it eventually fails short.  The electrons literally punch through the oxide layer.

 In view of the motor spec, that does sound like a more plausible reason for failure. Have you considered the possibility of a belt and braces solution, like fitting an additional hefty fast diode? Assuming that’s possible of course – I’m still not entirely clear about the motor topology here, and I’m assuming that this is basically a DC motor – is that correct?

[Les Jones 1Participant @lesjones1]

Hi Oliver,

               Have you checked the motor resistance between all commutator segments. What I mean is connect your meter across the motor, note the reading, turn the motor a few degrees and repeat the test. Do this until you have done one revolution. the reading should be about the same in any position. If it varies much then there is a short or open circuit on the windings. Either fault could kill the controller board. When the motor failed on my X3 two windings had shorted together and one winding went open circuit. I was lucky it just blew the fuse. You could also try running the motor from a 12 or 24 volt power supply. If  you do not have a current limited power supply you could use a car battery WITH A HEADLAMP BULB IN SERIES TO LIMIT THE CURRENT in case there is a short at one position.The motor should run smoothly (But at about 5% of its top speed rating.)

Les.

[Ian S CParticipant @iansc]

If theres a short the bulb will flicker as the motor rotates,were the short is the bulb will be brighter. Ian S C

[Steve GarnettParticipant @stevegarnett62550]

Well, it was the first post that made me wonder about something like a motor fault, I must say:

 

“I replaced the part with an IGBT that i had knocking around, and it worked OK, but the speed fluctuted under load until it finally failed.” 

 

The one thing that commutator damage could certainly lead to is speed fluctuations, after all… assuming that it’s some sort of bog-standard DC motor, which nobody yet has confirmed to me – I haven’t got any Seig kit at all, so wouldn’t really know how they were powered.

[Ian PParticipant @ianp]

As far as I know and assuming that all the ‘X1’ mills are the same then the motor is a permanent magnet brushed motor.

 

I found the motor/controller/and whole drive mechanism in general a bit lacking and have ditched the whole lot and converted my X1 to belt drive with a brushless servo motor (because I had to find some use for it!). I also modified the existing spindle to create a ‘quill’ arrangement.

 

Ian Phillips

 

[Les Jones 1Participant @lesjones1]

Hi Steve,

               I have assumed that the X1 has a permanent magnet DC motor similar to the X3.  The speed controller on the X3 uses two SCR’s and two diodes in a bridge which is phase controlled. I could not find a schematic on the web so I traced it out.

(If anyone wants a copy let me know.) The only Sieg speed controller schematic I could find was for the X2 This used two mosfets. My understanding is that Olivers controller has only one power device. I do not know how Oliver knew it was IGBT. It should be possible to test the controller when ALL the fault components have been replaced using a 100 watt light bulb as a load. (This works on the X3)

A warning to Oliver. If you think about looking at signals with a scope you will NOT be able to do so with it powered from the mains. THE CONTROL CIRCUIT “GROUND” WILL BE LIVE SO YOU CAN NOT CONNECT THE SCOPE GROUND TO IT. To test it safely you will need a floating supply.

Les

[Steve GarnettParticipant @stevegarnett62550]

Posted by Les Jones 1 on 09/02/2010 23:29:38:

Hi Steve,

               I have assumed that the X1 has a permanent magnet DC motor similar to the X3.  The speed controller on the X3 uses two SCR’s and two diodes in a bridge which is phase controlled.

 

Oh, that method… quite efficient, but potentially quite dangerous when it’s not isolated. But if the X3 uses the same control method, then just sticking one IGBT in – well, it’s not exactly going to work as intended, is it? So if there really aren’t two devices in it, the topology must be different.

 

If, on the other hand, there are two devices and they are thyristors (SCRS), then there’s half a chance that it would work if both devices were IGBTs – these are similar enough to SCRs to be a replacement device if the commutation conditions (and a few others) are met – and with an unsmoothed AC supply, some of them might be, I think. Even so, relying on the supply passing through zero isn’t the approved way to stop one conducting! I’d definitely be a bit worried about potential latch-up conditions, especially if an IGBT was put in to replace a device which had an intrinsic reverse-protection diode.

 

Bottom line – I think that the OP should establish exactly what sort of controller this is before sticking any more random devices in it!

 

Also, it’s worth mentioning that if you have a portable scope that’s battery powered, you can look at motor waveforms without any problems (no ground connection), but you’d probably want to keep your fingers away from it whilst doing so…

[Oliver LindleyParticipant @oliverlindley63277]

Hi all

 

Thanks for the further advice – i will check the motor with the meter or bulb method. 

I don’t think i made it clear above, but the mill worked fine for a couple of years with the IGBT fitted, it only recently became unstable speed wise, so i feel this is more likely the transistor failing than a problem with the motor.  The IGBT i removed was short across the collector/emitter.

 

Steve and Les – to answer your various questions…I did not know for sure it was an IGBT, hence my request for the actual part type at the beginning of this thread.  When the controller originally blew, i was in the middle of a job, and it was the only device of appropriate voltage that i had available – I did not know whether it would work until i tried it, but it did work absolutely fine, until it failed after years of use.

I was sure that the controller was PWM and not SCR based due to the drive components on the gate of the device, therefore, either an IGBT or MOSFET would do.  I did not have any appropriately rated diode to put across the IGBT, so I took my chances. 

The motor is 180V DC brushed type.

 

I have partially reverse engineered the controller and it seems that the mains input feeds directly into a bridge rectifier (no transient protection, EMC filter or PFC!).  The positive of the bridge feeds a current sense resistor 0.33R, followed by a fairly small toroidal inductor, through a relay arrangement (reversing and ‘UNNORMAL’ protection) into the motor.  The motor then connects to the MOSFET which returns to ground.

 

From this it is apparent that the “180V DC” is simply a function of the rectified mains (~340V = 240V x SQRT(2)) through the inductor/motor and the switching frequency of the PWM circuit.  When the IGBT failed, this would have exposed the full 340V to the motor and no doubt caused damage to other parts of the circuit as others have suggested.

 

I think my direction from here is to ditch the SIEG speed control and either build a proper one from scratch with a decent 180V DC supply for the motor, or as Ian P commented, ditch everything and replace with known good motor and controller.

 

I am an electronic engineer, and i design military electronics, so the commercial approach to doing things always concerns me (an occasionally impresses me) at how minimal and interdependant everything is.  The stuff we design has to be rated for the fault conditions, not just the operating ones, and then further derated to allow for reliability. 

[Steve GarnettParticipant @stevegarnett62550]

Posted by Oliver Lindley on 10/02/2010 23:33:00:

 

I am an electronic engineer, and i design military electronics, so the commercial approach to doing things always concerns me (an occasionally impresses me) at how minimal and interdependant everything is.  The stuff we design has to be rated for the fault conditions, not just the operating ones, and then further derated to allow for reliability. 

 

I think it’s just a bit unfair to lump the whole of commercial electronics together like that! We do the lot (because there are very few firms around now that can), and a lot of the stuff we do is commercial medical electronics – which has to be as near to 100% failsafe as you can ever get it. And we also designed a system that ensures that one particular mechanical device does fail safe – otherwise it could cause a lot of serious injury. And more recently, we updated part of the design for a project that we can’t talk about – suffice it to say that it has to be 100% failsafe under the most adverse conditions imaginable, and if this goes wrong, everybody will know about it…

 

I’d say that it was fair to describe cheap commercial electronics in the way you do – and that includes just about everything intended to be bought directly by the customer these days – unfortunately. A lot of people seem to like cheap Chinese lathes and mills because on the face of it, you get quite a lot for your money, and nowadays they seem to be improving. Maybe that’s why Les has found several different versions of what is essentially the same control system – they are improving it!

 

But an IGBT that fails after two years? In a machine with a brushed electric motor in it? If I was going to repair this at all, I think that firstly I’d be looking very carefully at that motor, and secondly I’d be putting a reverse-biased protection diode across the motor terminals – killing any reverse emf as close to source as you can. And then sticking another one across the regulating device. Chances are that originally this was a MOSFET device, and they have better switch-off characteristics than IGBTs, some of which don’t even have any sort of diode protection at all. If you were going to buy a new IGBT, then I’d definitely go for one specifically intended for motor control, because these have an extended Tsc (Short Circuit Withstand Time) and are less likely to blow up with inductive loads anyway. Something like this perhaps?

 

 

 

 

[Oliver LindleyParticipant @oliverlindley63277]

Steve, You are correct about the commercial electronics – I really meant el-cheapo Consumer  electronics.  There are plenty of non-military electronics where the standards are in fact higher.

 

I have got my mill going again – I have replaced the blown IGBT (50N6S2)  with another identical part, AND now a diode (RURG5060) across the IGBT.  Hopefully this will run until i get round to either placing a suitably sized order to Farnell for some of the IGBTs Steve suggested or the FETs mentioned earlier, or designing a better speed controller.

 

I took apart the C1 lathe we have in the workshop at work – which i wrongly assumed would be effectively the same as the X1 mill – and although the motor was the same, the control board was completely different.  I would imagine that Sieg are continually revising thier designs, putting right issues as they emerge.  We got the lathe last year, so it is considerably newer than my mill. 

 

Thanks to everyone for your input.

 

 

 

 

[John HaineParticipant @johnhaine32865]

Some time ago I bought a very nice DC servo motor which is destined to be fitted to my X1 when it gets CNC’d.  It came with a speed controller from KB Electronics in the USA – http://www.kbelectronics.com/ – it’s the KBWM type.  I haven’t tried it yet but I did see it operating at one of the shows.  These controllers use thyristors and are quite clever having speed compensation and have a good rep I think.  They have some UK stockists, not sure how much it would cost, could be a good replacement for one of these pesky X-series speed controls which judging from the X series forum are always blowing up.

[Steve GarnettParticipant @stevegarnett62550]

Posted by John Haine on 16/02/2010 21:23:35:
These controllers use thyristors and are quite clever having speed compensation and have a good rep I think. 

I’ve wondered about the wisdom of that before. The most (the only?) useful feature of a motor control that isn’t heavily compensated so that it stays at the selected speed until the opposing force defeats it is, of course, that it slows down under load. So you get a much clearer idea of when your increases in cutting force are really starting to bite. I think I’d be quite happy with compensated speed as long as there was an actual indication of the degree of compensation being applied, though – in fact that may be more use altogether than the uncompensated version. But I’d still want to see this issue addressed somehow, I think, and I do quite like the idea of a load meter!

[WALLACEParticipant @wallace]

Don’t know if it’s possible, but maybe loosing the belt a smidge so it slips under too heavy a load  might extend tje life of the IGBT’s !

 

As an aside, I built a 3 phase converter some years ago based on a design in Elektor magazine that was rated 750 watts  – but upped it by fitting bigger reservoir capacitors and a lot heaftyier IGBT’s.  As theyr’e voltage driven, fitting ones with bigger current capacity worked in practice – I chose to ignore (i.e. didn’t know enough to make a judgement  !)  variations in input capacitance, switching speed etc. 

 

I have blown up the odd one since   –  usually a complete phase goes down and the inverter won’t work at all as the current limiter comes in and kills the drive to all the IGBT’s.

 

WALLACE.  

[Author]

Posts