Murex Autolynx 2 no weld power, work in progress

[mike 109444]

Have recently purchasedan BOC Murex Autolynx 2 as a non working machine. Machine looked to be in tidy condition but has the fault of “no welding output “ I had thought there would be plenty of info on the forum on this machine but it seams that anyone that's posted for info have not stuck around after (fixing?) there machine following getting Dave @NotANormalCoder to assist !


Did try a PM to Dave but not heard anything back yet. I have therefore carried on checking through the machine trying to figure out the fault(s).


So far :

1. Hernia moving it !
2. Have sorted out some poor connections on the multi pin plugs that connect to the pcbs as these were causing wire feed etc to dropout when wriggled with torch trigger pressed.
   
3. Re-flowed some then all! Solder joints on the three pcb's
   
4. Checked and tightened all the tag board strip connections (TB's)
   
5. Checked and tightened the spade connections on the two link wires that feed the SCR.
   
6. Checked for continuity from the mains input up to and through the on/off switch to the TB connections of TB4 where it then feeds onto several areas of the machine/pcb's (still working through / working out what is feed from this).
   
7. Have previously done basic diode and turn on checks on the SCR but as I have not had to work with these devices in my normal job (they may have been present in psu's I have replaced but not needed to repair any) so have looked up how these devices are used and work. Not sure that the one in these welders is used in quite the way I have read up on apart for maybe switching on at zero crossing ?
   
8. Noted that the SCR has a bridging strip of brass (aprox 20mm x 10mm x 2mm !) between connection 2 and 3. I have no ideal why this is !
   
9. Have gone over the three pcb's identifying nearly all the semiconductor devices. (one transistor metal can pitted/ non readable) These include an op amp, bipolar trigger ic for triacs, quad input nand gate, atl east three 555 timers and many other devices.
   
10. Have checked operation of any of the relays that are being feed power to there coil and all seam ok.
    
11. Disconnected the output diodes on the heat sink and checked for shorting and basic diode checks, all ok.
    
12. Some of the checks were done to avoid trying to fix a machine with an expensive faulty part that would rule out continuing the repairing of this machine.

Thoughts on thismachine so far. Machine is well built (like a tank!) Not a lot ofroom to work in it. Wiring loom very tightly bundled making following/tracing wires difficult.





Wish I had looked upthe weight of this machine before hand (seen quoted as 98Kg's)


Glad the chap I boughtit from was there to give me a hand humping it in to the back of myestate car. Would have helped if it did not still have a near full15Kg spool of wire in it !!





Only time so far I was glad that the new Skoda Octavia (hybrid) (company car) has a flat load area in theback because for work the loss of boot space due to the batteries isan absolute PITA !

 

[eddie49]

That looks like a good purchase, in really nice clean condition.
I have liked these welders ever since I looked inside one at a machinery auction three decades ago but was outbid on it, at about £240.
I am ashamed to admit that I have two of them, and neither of them work. One belongs to a friend - my "MOT Man", a garage owner, and I undertook to fix it for him over a year ago. The other I bought "partially working" from a returning American expatriate 28 years ago. That one even came with a manual, which I have misplaced.....

Regarding your point number 7 : the SCR ( s ?? ) may be a "light dimmer" on the primary of the main welding transformer, or, if large enough, it/they may be part of a controlled bridge rectifier on the low-voltage high-current secondary. Either way, the key is to trace the trigger wires from the gate(s) back to a pulse circuit on one of the PCBs. To provide electrical isolation, this may involve pulse transformers, as in this PowCon thread:

Powcon the Legend - 300 SM, low OCV

I have a powcon 300 SM that has low OCV, is only reading 28 volts with no load, but I have 340 volts across both main capacitors and each side of my "filter board" or rectifier board is putting out 150 volts. However, I have no ticking sound. The screen also reads "HHH" on the LCD. Anyone...

www.mig-welding.co.uk

I will try searching for my Autolynx manual.

 

[mike 109444]

Hi Eddie
Yes the machine does look to be in very clean condition dispite previous owner being a farmer ! (No offence to our farmer's of the UK, keep up the good work lads n lasses)
I refered to the device as a SCR as that is what it has been refered to in older post and looking up the device (for testing info) it seams to be refered to by several different names. It is a semipack 1.
If you look at the 1st image i have attached of this device you may be able to see a black bar, this is the edge of the brass strip that is linking the top two connections togeather ! (the + and -). I have isolated the devices connections and done the "metering as a diode" then triggering the gates one at a time tests and the device works as expected. I belive that the machine does use control of this device to vary the duty cycle (?) of the transformers, there are 5 transformers in the base! One is I belive is for the low voltage for pbc logic and relays. One I would have thought is an inductor for the DC weld output leaving 3 for the welding voltage though it is difficult to see how these are all connected and used due to the compactness of this area.
From Dave's comments on older post for these machines saying "you would need a scope to do some of the settings" I would think that is to set up the triggering of the the gate.
The 2nd image is of the pcb that the two gate wires connect to. They connect top corner indicated by the G1 and G2 writen on pcb.
If you do get a chance to dig out any docs for this machine it would be much appreciated as there is next to none on the net.
Will take a look at the link you posted.

 

[eddie49]

The brass bar joins the cathode of thyristor 1 ( + ) to the anode of thyristor 2 ( - ). Since the anode of 1 is already linked to the cathode of 2 internally in the Semikron Semipack 1, the two thyristors are wired back-to-back ( "anti-parallel " ), making them into the equivalent of a TRIAC and giving full control of both the positive and the negative half-cycles of the AC input.
The size of the thyristor module and the gauge of the wires means that it is rated in tens of Amps, not hundreds, so it must be in the primary circuit, varying the welding output by phase control of the AC input power.
The TCA280A integrated circuit generates trigger pulses for the thyristors a variable time after the zero-crossing point of the AC waveform. I think the pulses are amplified by the TO-5 metal-can transistor on the left, and coupled to the gates by the 78411 ferrite-toroid pulse transformer above it. I think there will be one primary and two identical secondaries on that transformer. You could use an oscilloscope on the primary to check for the existence of those pulses.

 

[mike 109444]

Thanks for deciphering that Eddie. I will have to dig out and dust off my scope! I think first I will maybe check over that pcb for basic supply to the devices.
Just been giving myself a basic course in scr's on YT. Most seam to be from the non English speeking parts of the world !!
One I did listen to was also discribing testing of the "puck" type mentioned in you previous link and (FYI) noted the guy doing the test said that to be able to test these you should apply a small amount of presure to the top and bottom (A & K) as that is how the device is used when installed. He used a pair of mole grips with one jaw wraped in insulation tape to very lightly apply pressure before testing. Worth noting.
ps 3:08 AM posting ! You get up early

 

[mike 109444]

Continuing on
Well I have checked thepower across the board 3 (pcb has the 78411 transformer on it).

Supply to the CA748 is15.6v.
Supply to TCA280 is15.6v.

There is no AC feedinginto the small bridge rectifier (bottom corner) This Ac comes on tothe pcb via the PL302 pins 9 & 10 from a small transformer justbehind the Semipack. This transformer meters ok on resistance of it'sinput and out put. Feed to this Transformer comes up from the TB2where several of the large transformers are connected but not surewhat's going on there.
There is a very small7815 reg on pcb that is out putting 15v ok.

Measured the resistanceacross the pins of the 78411 to work out pin config and of the pinsbeing used in circuit:
2 to 3 = 2.5 Ohms
4 to 5 = 1.4 Ohms
6 to 7 = 1.5 Ohms.

With machine powered up:
TAC280 IC
pin 1 showed a roundedoff top and bottom square wave. Can't recall voltage but believe wasquite high (near mains ?).

pin 2 showed a sawtooth waveform.

Pin 3 showed on DVM2.2vdc

pin 4 n/c

pin 5 showed on DVM 5.5changed to 1.1 when trigger pulled

Pin 6 showed a sawtooth waveform.**

Pin 7 showed on DVM 0changed to 15 when trigger pulled

pin 8 Showed on DVM 15v constant.


Pin 9 showed on DVM15.3 changed to 14.6 whentrigger pulled

pin 10 showed on DVM 0changed to 1.4 when triggerpulled.

Pin 11 + Supply to chip15.6v

pin 12 n/c

pin 13 n/c

pin 14 showed on DVM 0changed to 5 when trigger pulled.

Pin 15 showed on DVM15.5 changed to 14.2 when trigger pulled.

Pin 16 0v supply.


** There were waveforms on:


Pin 6 saw toothwaveform


pin 14 nearest Idescribe as positive ramp wave form.


Pin 15 nearest Idescribe as positive ramp wave form with more of a flipped capitalletter L !

CA784 IC
pin 1 n/c

pin 2 13v

pin 3 1.1v

pin 4 0v supply

pin 5 n/c

pin 6 2.1v

pin 7 + supply to chip15v

pin 8 n/c


Readings taken from the78411 to PL 302 pins:

Pins 2 & 3 20v peekto peek rounded top sine wave.

Pins 4 & 5 240vpeek to peek, flattened off peeks sine wave.

Pins 6 & 7 1 v peekto peek choppy sine wave !
Any further thoughts Eddie ?

 

[mike 109444]

ps it's been atleast 20 years since I last used a scope and even then I would be doing comparisons with known or expected wave forms (usually digital or radio) which is a bit different than trouble shooting on kit you have never needed to scope before . So my apologies to those that regulary use a scope and wince at my discriptions of wave forms

The above test were done with a Hameg HM203-5 last calibrated about 1988
And a Lidl dvm that was never calibrated

 

[Colin Barron]

You are lucky that you have the later version using SEMIPAC controllers, my first one had i think triac which kept blowing. I ended up buying another one with a semipac which was pretty rough and making a good one out of the pair. I never did find out why the earlier model kept blowing but a welder repair man said the earlier model was not as reliable. Check the capacitors they are now over 30 years old.

 

[mike 109444]

Hi Colin
I was unaware there was an early verses later version !
From Eddies explination of the strange way that the Semipack is configured in mine, with a strap (brass) strip between the number 2 and 3 connections it works as a Triac. I recently had to swapout the cap in my other welder, just as well I had a spare as price for large caps get to the point of non economical repair. Hope the ones in this mc are good cos as it has 3 that could be costly.
Might have been with your orignal machine that an up rated device may have cured it but you need to first know what the orignal device id is! (there are no numbers on the one in this machine)
I must admit I am struggling with this machine due to the very different way that the power runs through / is controled in the machine and with no schmatics or any machine that is similar to work it out from. Still as most components in the machine are off the shelf, will hopefully get there in the end.

 

[mike 109444]

Hi @eddie49 not sure if you have had a chance to look at my post number 6. Any thoughts ?

 

[eddie49]

Regarding the SEMIPACK ( i.e. 2 x thyristors ) vs. TRIAC upgrade:
Both of these types of electronic switch have a hard time dealing with large inductive loads ( e.g. transformers ). This is mostly because when they are used to provide stepless control of power levels, known as phase control, the device must stop conducting at the end of every half-cycle, ready to be triggered again some time after the start of the next half-cycle. With an inductive load, the current lags the voltage, so when the AC voltage drops to zero at the end of a half-cycle ( the "zero-crossing point" ), current is still flowing. Synchronisation between the trigger circuit and the device is lost. This can lead to high unidirectional current flowing, which destroys the device. Apparently due to the way the semiconductor layers are created, thyristors are more resistant to this problem than TRIACs.

The resistance values for the coils on the ferrite-ring pulse transformer imply that the two equal ones ( pins 4 & 5, 1.4 Ohms and pins 6 & 7, 1.5 Ohms ) are the secondaries used to trigger the thyristor gates, and the coil across pins 2 & 3 is the primary. The primary is probably driven by a transistor, fed from the output pin 10 on the TCA280A pulse generator chip. Sending pulses out from this pin 10 a variable time after the zero-crossing point of every half-cycle the AC mains supply is the primary function of the TCA280A. Those pulses are supposed to be 2.8 volts less than the chip supply voltage, which is 15.6v on pin 11. In fact pin 10 showed 1.4v, but that was measured on a DVM, rather than with the 'scope.

If the ratio of the coil resistances equals the turns ratios ( i.e. the same diameter wire is used for all of the coils ), then the 78411 is a step-down transformer, with ratios about 2:1:1. From the 20 volts measured on the primary, the gate voltage should be about 10v, which is on the high side but still possible. However, the voltages and waveform shapes observed are quite different. The gate trigger should be a sharp pulse, only 100 microseconds long.

I have not found my Autolynx manual, and am beginning to doubt that I had it. I should get one or both of the Autolynx 2s out of storage, try powering them up, and make some measurements.....

When you power-up the welder using the rotary control panel switch in any of the three modes, does anything click inside?
Is there any relay, solenoid or motor operating when you squeeze the torch trigger?
Any change if you gently tap or press wires, terminal strips, plugs, or the PCBs with an insulated tool whilst powered-up and triggered?
Not knowing which transformer does what makes it difficult, but you could check for current flow through the Semipack by inserting a 100W filament lamp ( or an AC ammeter ) in series with it ( at the push-on terminal tagstrip below the Semipack, then vary the Volts control ).

 

[mike 109444]

A short reply for now Eddie. I have the machine upn and running . One fault caued possible by me...and two components that were duff (probbley orignal fault) I found that the resistor R 337 (just above the bipolar tranny) was reading 400 odd ohms and it's a 17 ohm one. Replaced that which then promptly smoked ! So checked what it was connected to and found that the tantalum cap C 313 wah shorted. Replaced cap and resistor and still not working. So checked for pulses from pin 2 of the TCA 280 as these should feed into pin 6 and the comparison made should output on pin 7. I was at first getting pulses at the top of R321 but very little if anything from the bottom of R321. Then sometimes nothing at top of R321 ! (kept thinking poor scope probe connection) Then found that with no pulses at top of R321 there were pulses at the chip pin 2. Took pcb back out and found that the pcb track that the R321should be attached to was lifting/broken, so pulses not getting through R321 AND not feeding along the track to other resistors etc. Once I repaired (my caused) damage pulses feeding through and welder outputting welding voltage.
The attached pdf from way back helped with some of the explination of the TCA 280 inparticular section 2 basic operation.
I will try and get some sketches of the scope outputs from the various test points when time and work allows to update this thread for any one following on.
Thank you Eddie for spending the time investigating the circuits and explaining them to me.
I was starting to think the TCA 280 might be duff and when looking around for a replacement they are rare and expensive as a comment on one forum did say they were superceeded not long after introduction due to being "not a good devlopment" and replaced by the TCA780 and TCA 785 but not as a pin for pin replacement.
Hope you dig out your machine and we can maybe work through it
Thanls again