Oscilloscope grounding confusion

[Uncle Bulgaria]

I've got myself into a muddle, and my oscilloscope manual is not making it clear to me. Reading threads online has just increased my confusion.

If I have a mains powered circuit, with an opamp group run from a secondary tap via a bridge rectifier, where do I put the probe ground to monitor the opamp outputs?

The opamps' power supply is +-12V with a virtual ground. The ground is referenced to the opamps' output in the end - a thyristor gate drive circuit for a welder's current control.

Do I ground the probe to the chassis and earth ground, the virtual 0V or the -12V rail which is the 0V of the bridge rectifier?

[John_BS]

Can you post the circuit? Devil in the detail!

John

[Uncle Bulgaria]

Sure, John. It's attached. I'm afraid it runs over two badly-scanned pages in the PDF manual I found, so the middle components are half-missing, but legible enough. I've combined them into one page.

[Herald1360]

Hmm, there doesn't appear to be any secondary side connection to earth unless the workpiece happens to be grounded in which case all bets are off!

In principle, you could attach your scope ground anywhere but.......

Has your scope got two channels with the ability to add/invert so you can use it in single channel differential mode? No need to use the probe earths at all in that mode. Are you lucky enough to have a battery portable scope with isolated inputs in which case you can do what you like with the probe "earth" connections. Ditto if you have an isolation amplifier type probe.

Good luck!

[Keith956]

I would have said the rail marked 0V is what you connect your scope ground to. You've got isolation from the AC mains by way of the transformer, so anywhere on the output side can be safely grounded, and the 0V rail is what the designer has decided to call ground.

[Terry_VK5TM]

The 0v line is the common reference "gnd" point for the op-amp's.

The little diagram of the op-amp ic shows the power pins connected to + & -12 V respectively and the junction of ZD1, ZD2, C5 & C6 form the gnd point.

[Radio Wrangler]

This is easy.

It's a DC welder.

The whole of the welding side is isolated from mains and is isolated from ground. This is how it should be.

Some welding processes are done with the electrode positive and the work negative - relative to each other. The electrode is insulated and is treated as 'live', the work probably isn't and is treated as ground. The work isn't grounded by the welding machine, but may be via its environment. The reason for this polarity is that electrons are real, holes are lacks of electrons where one might have been. Heat dissipation is uneven, electron bombardment heats the electrode strongly in addition to the heat from current flow. THe work gets current flow heating but misses out on electron bombardment, in fact electron emission consumes energy. The electrons emitted by a cathode cool it a little. If you're doing coated electrode 'stick' welding the heat on the electrode is needed to melt the flux, turn it reactive and release a cloud of CO2 to shield the molten metal from atmospheric oxygen. The spray of molten metal from the electrode conveys heat to the work. So electrode positive is normal for the most common electric welding process of the past century.

However, some welding processes are used with the electrode negative, even some coated stick electrodes are made that need this polarity. In this case the work gets the lion's share of the heat and this helps the weld pool penetrate more deeply into the work, making a much stronger joint. The electrode has to have a coating that works with the (relatively) lower heat input.

So, welders are made with floating outputs so the guy doing the welding can choose to connect his electrode holder and his work clamp to whichever terminal suits the rod clamped in the holder. and either could go to planet earth, though it makes more sense for the work to be earthy.

So! is you want to go scoping around on the secondary sides of this machine, you can put the scope ground clip anywhere you want, in theory, just so long as you don't make any other ground anywhere else.

I said in theory. If the place you want to ground is in the midst of an inverter (this unit is not an inverter) it could have a strong high frequency content, and the stray capacitance of things like the transformer could give you a problem with current in your scope probe ground clip. It should be OK with this unit.

This welder has a big iron transformer and does current control by controlling the phase of triggering those big thyristors in the rectifier bridge. That's what the opamp-ery is for.

it senses welding current to decide when to net fire the SCRs. So to see the thing working, you'll need some welding current. This is going to be difficult. You want a dummy load that will dissipate something like 1kW at 50 amps if the welder is not wound up much. As you're fixing the current control, it might wind up flat out and the dummy load would need to handle several kW.

Sorting out a suitable dummy load is the hardest part of fixing these things, and for scoping around, the dummy load has to float.

Plan B is to check the transformer alone, try the two big diodes, and check the thyristors aren't shorted on. After that, all the parts are cheap... you could just replace the lot.

So, no, putting the scope probe ground clip on the chassis is pointless. It is isolated from the circuitry you are going to look at.

David

[trobbins]

As indicated there should be no issue with leaving the secondary floating, for the purposes of scope probing. Perhaps for starters, remove the thyristor gate terminals and energise the primary and confirm all the control circuitry is operating ok, including using a variable mV supply to replace the shunt signal to confirm current control.

The mains inrush-may be significant. If you can set current control to min, and include a suitable load (and a shorted output should be ok), you can monitor the work output voltage waveform for waveform symmetry to confirm each thyristor is ok, but you will need sufficient load to allow the thryistors to turn on. If the unit is old then it may be worth inspecting each thyristor and diode and heatsink for good thermal contact without indications of pitting or other interface thermal stress, and also any e-cap, and either check or replace the secondary side MOV.

[Radio Wrangler]

By the way, Murex was founded as the equipment arm of BOC (British Oxygen Company, now part of the German Linde group) Their equipment was highly reputable and rather expensive. Some firm paid a lot of money for that welder. It's well worth fixing so long as the transformer is OK.

Current (bad pun!) DC welders have almost all moved over onto inverter architectures with ferrite high frequency transformers and have lost a lot of weight. UB's welder is halfway back to old school technology which did current control with a flux shunt in the transformer.

David

[Uncle Bulgaria]

Easy for you perhaps, David! For we bears of little brain a little knowledge is rather a confusing thing.

It's probably that I can't get my head around isolation transformers. If earth is the chassis then my brain boils trying to see why it matters that the live is galvanically isolated from the real live. There are still volts and current in the circuit, and electrically the earth is the same point for primary and secondary.

I have left off a dummy load because I thought the circuit sums the firing impulses, and if the component from the output is zero, there'll still be the angle derived from the ramp generator to check whether the opamps are working.

The welder was working last week, and I went to finish off a job and got no arc. It's just been turned off and left inside, so I'm confused. It is a BOC branded machine, and has the ability to make my arc welding look competent so I'm keen to fix it.

The 'scopes available are a Cro-Tech 3131 (thanks to Diabolical Artificer!) and a Tektronix 2445.

[Herald1360]

The physical metalwork chassis of the welder is connected to mains earth via the mains input lead for electrical safety reasons. The isolating step-down transformer secondary side circuits (both welding and auxiliary) are floating (ie not connected to earth or any other reference level) unless one or other of the welding output connections is earthed via an incidentally earthed workpiece.


Your 2445 'scope does have the add/invert functionality (see p 3.3 of operator's manual) so I would suggest you use this to be on the safe side (from the 'scope's point of view). Leave the probe grounds disconnected (unclip the leads or make sure the clips dont touch anything) and use probe 2 connected to the 0V mid rail as your reference and probe 1 to look at opamp signals etc.


Have fun and good luck!

[Radio Wrangler]

There is a heavy current shunt "SH1" in the connection from the +ve output of the big bridge to L1 nd hence the + output terminal. THis is used to sense the welding current. The voltage drop on the shunt is buffered and amplified by IC1

This gets combined with the voltage from the current control pot RV1 by IC2. IC2 has a biggish (for opamp circuits) capacitor and R8 in its feedback network to slow it down to average out the pulse-y nature of the rectified but unsmoothed welding current.

The output of U2 is the command for how hard to try, think of it as the throttle on an engine with a governor. IC3 must be timing ramp generator... not clear to see given the mismatch in the joined diagrams. IC4 compares the 'throttle' voltage with the ramp voltage and decides when they cross. It then flips its output positive making a pulse through C3 and TR1 acts as a emitter follower to boost the currentR15 and R16 share the current between the two big thyristors.

The thyristors only turn off at the zero crossing of the mains waveform. They turn themselves off. The circuitry only has to decide when to turn them on. Control is established by how much before the mains zero crossing the thyristors are triggered. The sooner it is done, the longer the SCRs are on and the greater the Oomph.

The point labelled 0v would be a good point for your earth clip. You'd think someone had been this way before....


Note that with no load on the output, IC1 says I see no current.

The current pot says there should be X amps!

IC2 says increase the throttle!

And the welder tries to run flat out.

You may hit trouble with the thyristor trigger trying to be too soon. So some load helps.

David

[MotorBikeLes]

I acquired one of these welders 10 years ago. When I came to try it out, I found the mains transformer primary was shorting to ground. I did consider mounting the transformer on insulators and removing earth, but in the end decided to scrap it. The carcase with transformer still outside ready to be weighed in. I likely have all the remaining bits, but not sure. I could have thrown the board, but thyristors probably still around.
Les.

[Uncle Bulgaria]

Attached are the opamp output traces at x10. As David explains that without SH1 dropping voltage the circuit will be full-on trying to catch up, I had hoped that the opamps would demonstrate this, load or no.

As they aren't showing much of anything, perhaps it's the triggering after all. I haven't tested the thyristors yet as they're very awkward of access. If it is them, I'd be glad of your spares, MotorbikeLes! I don't know why either would suddenly stop working.

IC numbering follows the Murex inset diagram, rather than the CA324E datasheet. Probe grounded at 0V between the Zener diodes.

All at x10 attenuation.
IC1: 2mV/div, 5ms/div.
IC2: 2mV/div, 10ms/div.
IC3: 0.5V/div, 2ms/div.
IC4: 2mV/div, 10ms/div.

So there's a 15V pulse of peculiar shape on IC3, but not much of anything else. Yes, IC3 is the ramp circuit. There doesn't seem to be anything missing from the scan, just D9 and R14 from the +bridge to ZD1 are peeping onto the upper half. There's a 'circuit description' in the full manual, but the circuit diagram's more difficult to read.

[trobbins]

Does the current shunt amplifier give an output that is proportional to the shunt voltage? If you read the shunt FS voltage (eg. 100mV) and make up a 100mV calibration supply (eg. from a battery and a pot) and connect that instead of the shunt, then does the opamp output provide a suitable output as shunt voltage is increased from 0mV to FS mV?

[Uncle Bulgaria]

Good idea, trobbins. Output was still nada, however. After going round the houses with 1.5hrs on hold to UPS whose operatives appear to have a hazy grasp of English, they finally put me onto the Bodmin depot where a competent person who agreed that my parcel from CPC didn't need an 'apartment number' personally fished it out of the 'return to sender' pile.

It contained an enormous amount of fresh air (why is there an inverse relationship between object and parcel volume?) but crucially some LM324AN ICs. On replacing with one of these, I get 40V on the electrode outputs so I'm going to see whether it arcs now.

The machine turns on as soon as it's plugged in, as there's no on/off switch, so it is always running the thyristors at the pulse frequency selected with the current potentiometer. This feels like it would have some effect on life-expectancy, but as the thing's obviously designed as a workhorse and is still working it can't have too much effect.

[Uncle Bulgaria]

Objective Complete. I've run some lovely beads with it, so it appears the IC was the only bust component. Thank you for all the help. I feel a bit more 'scope-confident now, too.

[Radio Wrangler]

That's good. A decent success story is always cheering.

On the scope front, you'll see that seasoned engineers reach for their scope not their multimeter as their first have-a-look machine. You don't get the precision of a meter, but you get an awful lot more information and clues. It turns out you don't often need the precision, but the more clues the merrier.

David

[Al (astral highway)]

Originally Posted by Uncle Bulgaria

Objective Complete. I've run some lovely beads with it, so it appears the IC was the only bust component.

Happy days! I'm glad your tenacity paid off and you fixed it ok, UncleB