Brother industrial sewing machine repair / reverse engineering (repair success)

[max.wwwang]

[Edit]
This is a thread on my repair and reverse engineering project with a Brother industrial sewing machine.
It started off with a question "Why does the polarity of oscilloscope probe matter?" then touches on mains earthing protection (or grounding, depending on where you live) in some good depth, then aspects of the project. And it goes on...
[/Edit]

I'm repairing another vintage industrial sewing machine (the other one is here). I found something that I can't find an explanation for so would appreciate any help from you.

I'm aware of the risk of blowing a piece of equipment by using an oscilloscope in the wrong way (particularly when using two probes), after having watched the vids from David. So I would not put the Earth clip of the probe to anywhere that is not connected to Earth in a circuit lightly. In this case however, it appears Earth is not connected to any point of the circuit, so it's safe to clamp the clip to anywhere I want as long as I'm using only one probe.

I'm testing the output terminals intended to drive a solenoid. The circuit is a typical design, consisting of darlingtons, flyback diodes and associated power resistors.

Without the load (solenoid) connected, when connecting the probe in one way, I can see a switching between ~40V DC and 0V, as a waveform on the oscilloscope, across the two pins. When I change the polarity of the probe to another way, the wave form becomes a flat line.

What's going on here? Thanks for any insight.

[Psi]

Without the solenoid connected there may not be anything to switch on/off and so you might be measuring something that is now floating.  Are you sure it's a push/pull driver that will function without the solenoid in circuit?

Also keep in mine its not just earth connections you have to be aware of when deciding where it's safe to clip the scope ground.
Neutral is connected to Earth at the house fuse box, so electrically they are the same wire.
And of course phase has voltage with respect to both those as well.

If the device has a transformer and you can see it's totally isolated from mains and the case is not earthed then it's probably fine to clip anywhere with a single scope probe.

[max.wwwang]

Without the solenoid connected there may not be anything to switch on/off and so you might be measuring something that is now floating.  Are you sure it's a push/pull driver that will function without the solenoid in circuit?

It's not a push/pull circuit or something like that. It's a digital circuit beautifully made without an MCU chip but with only logic gates (flip-flops). The existence of a waveform, with only the power on (but without any action on the pedal lever), is not expected and is probably a problem on its own. But now what baffles me is the different behaviour of the oscilloscope with different ways of probing the two pins.

Yes, they are probably floating. But, if I understand correctly, notwithstanding the floating pins, the relative voltage difference should be correctly reflected between the two poles of the probe. And clipping the probe to one of the two terminals will set it to the ground potential anyway. Am I right here?

And true, the behaviour is different when the load is connected (again no voltage differences or waveforms). I'm not sure if that is due to the existence of the flyback diode?

[max.wwwang]

Neutral is connected to Earth at the house fuse box, so electrically they are the same wire.

Are you sure about this? If this is the case, I will detect a positive connectivity with a multimeter between ground and neutral of a wall socket. Really?

Edit: just checked this out - indeed neutral and ground of the wall socket are electrically connected. I still have not truly understood how ground protection of the grid works. 

Any recommendation of good resources about this? I watched videos of TPAI on YouTube, thought I understood, but obviously I have not.

The good side of this is, this makes me understand (really?) why an isolation transformer is useful/important in repair jobs (I made one at one point by the way, though have not really used it).

[srb1954]

I'm repairing another vintage industrial sewing machine (the other one is here). I found something that I can't find an explanation for so would appreciate any help from you.

I'm aware of the risk of blowing a piece of equipment by using an oscilloscope in the wrong way (particularly when using two probes), after having watched the vids from David. So I would not put the Earth clip of the probe to anywhere that is not connected to Earth in a circuit lightly. In this case however, it appears Earth is not connected to any point of the circuit, so it's safe to clamp the clip to anywhere I want as long as I'm using only one probe.

I'm testing the output terminals intended to drive a solenoid. The circuit is a typical design, consisting of darlingtons, flyback diodes and associated power resistors.

Without the load (solenoid) connected, when connecting the probe in one way, I can see a switching between ~40V DC and 0V, as a waveform on the oscilloscope, across the two pins. When I change the polarity of the probe to another way, the wave form becomes a flat line.

What's going on here? Thanks for any insight.

The 2SD720 transistor cases, which are also the collector connections, don't appear to be insulated from their respective heat sinks so you should assume that these heat sinks are not grounded and you can't clip your scope earth lead onto them. Remember, for a standard scope, the earth clip is connected through to the mains earth connection. Clipping a scope lead to such a heat sink will likely short out an active signal.

Assuming that your device has a mains power transformer it is best to connect your earth clip to the circuit common, which will be found on pin 7 of a 14-pin logic chip or pin 8 of a 16-pin chip. The heat sink of the regulator (the vertical one on the RHS) is also likely connected to circuit common (check with an ohmmeter) and could be safely clipped onto.

If your device doesn't have a mains transformer then you will need to power it through an isolation transformer before you can safely connect your scope earth lead to the circuit common. 

[james_s]

Neutral is connected to Earth at the house fuse box, so electrically they are the same wire.

Are you sure about this? If this is the case, I will detect a positive connectivity with a multimeter between ground and neutral of a wall socket. Really?

Edit: just checked this out - indeed neutral and ground of the wall socket are electrically connected. I still have not truly understood how ground protection of the grid works. 

Of course they are. Every electrical system that I'm aware of around the world has neutral tied to earth somewhere. Here it's in the main panel, elsewhere it may be somewhere else, but it's always done somewhere. Neutral is a current carrying conductor, a load on a circuit may pull it up above ground at the point of use, sometimes as high as a few volts due to resistance of the wire. The protective earth ground is for safety only, it should never be carrying any current except in a fault condition, if a live wire contacts the housing of a metal appliance for example if it is grounded that will cause a short circuit and trip the breaker and prevent the housing from becoming live.

If you're not 100% sure you understand all this, you shouldn't be messing around with non-isolated gear (such as a scope) and anything that is mains connected.

[max.wwwang]

The 2SD720 transistor cases, which are also the collector connections, don't appear to be insulated from their respective heat sinks so you should assume that these heat sinks are not grounded and you can't clip your scope earth lead onto them. Remember, for a standard scope, the earth clip is connected through to the mains earth connection. Clipping a scope lead to such a heat sink will likely short out an active signal.

Thanks! Yes, the case of 2SD720 is not isolated from the heatsink. That's why I clip the probe on it (it is connected to one pin of the output). Yes, I understand one pole of the probe (the little clip on the side) is directly connected to mains ground. That's why I was careful and checked that mains gound is not connected to any point of the circuit.

Assuming that your device has a mains power transformer it is best to connect your earth clip to the circuit common, which will be found on pin 7 of a 14-pin logic chip or pin 8 of a 16-pin chip. The heat sink of the regulator (the vertical one on the RHS) is also likely connected to circuit common (check with an ohmmeter) and could be safely clipped onto.

If your device doesn't have a mains transformer then you will need to power it through an isolation transformer before you can safely connect your scope earth lead to the circuit common.

One good thing is that I'm using a mains transformer to mock the power supply to this unit (30V AC), which is isolated. So I have not (hopefully) blown it by probing it with the scope.

What looks like a voltage regulator (the one you refer to) is not a regulator, only another darlington (D633).

[IanB]

Of course they are. Every electrical system that I'm aware of around the world has neutral tied to earth somewhere. Here it's in the main panel, elsewhere it may be somewhere else, but it's always done somewhere.

This is a digression, but not exactly true. There are some places that use the IT earthing system, where none of the live conductors are tied to earth. This is found quite commonly in Norway, for example. On the other hand, this means there is no "neutral" conductor, so in a way your statement is true. If there is no neutral, it cannot be tied to earth.

[max.wwwang]

Of course they are. Every electrical system that I'm aware of around the world has neutral tied to earth somewhere. Here it's in the main panel, elsewhere it may be somewhere else, but it's always done somewhere. Neutral is a current carrying conductor, a load on a circuit may pull it up above ground at the point of use, sometimes as high as a few volts due to resistance of the wire. The protective earth ground is for safety only, it should never be carrying any current except in a fault condition, if a live wire contacts the housing of a metal appliance for example if it is grounded that will cause a short circuit and trip the breaker and prevent the housing from becoming live.

If you're not 100% sure you understand all this, you shouldn't be messing around with non-isolated gear (such as a scope) and anything that is mains connected.

How dumb I was!  But anyway now I know (a part of it)!

Just to understand this better ( ), is it the only reason for connecting neutral to ground (say, in the switch box) that the breaker will trip when a live wire touches the grounded metal case of an appliance? Certainly will do a bit more homework on this.

[james_s]

How dumb I was!  But anyway now I know (a part of it)!

Just to understand this better ( ), is it the only reason for connecting neutral to ground (say, in the switch box) that the breaker will trip when a live wire touches the grounded metal case of an appliance? Certainly will do a bit more homework on this.

There are multiple reasons to reference things to earth. One reason is common mode voltage. If the transformer feeding your house didn't have the secondary bonded to earth, it's possible you could get a high common mode voltage due either to capacitive coupling, or worse a leakage or short. In other words live and neutral could both be floating at 7200V or whatever feeds the primary of the distribution transformer. This is as you might imagine, a very bad situation.

[tautech]

How dumb I was!  But anyway now I know (a part of it)!

Just to understand this better ( ), is it the only reason for connecting neutral to ground (say, in the switch box) that the breaker will trip when a live wire touches the grounded metal case of an appliance? Certainly will do a bit more homework on this.

There are multiple reasons to reference things to earth. One reason is common mode voltage. If the transformer feeding your house didn't have the secondary bonded to earth, it's possible you could get a high common mode voltage due either to capacitive coupling, or worse a leakage or short. In other words live and neutral could both be floating at 7200V or whatever feeds the primary of the distribution transformer. This is as you might imagine, a very bad situation.

11kV in NZ is normal for last mile HV normally stepped down from 33kV used for medium rage distribution and it derived from 110kV or 220kV, our normal long range national grid voltages albeit we have some 400kV main trunk feeders from our hydro regions also.

NZ and AU have a quite similar mains distribution network and share the same single phase socket for 230VAC mains and our 3 phase supplies are 415VAC normally with only an earth wire although some machinery that uses both 3 and single phase require a 5 core cable as it's not legal to use the PE as a neutral return despite them being bonded at the fuse/metered supply.
Each 415VAC phase has that potential only to any of its accompanying 2 phases and they each are at 230VAC to mains ground/earth.
2 phase supplies are quite rare here although it is sometimes used into high load domestic applications however a 3 phase supply is more common and good electricians can take some trouble to spread these 3 single phase loads onto each of the phases to keep with limits of our normal 60A max/phase supplies.
Uprating a supply, even for just one phase adds significant cost to the daily line charges in NZ.

For a rough idea, per phase we pay a little over $1/day with kWh rates levied on top for a 60A single phase supply.

For some special applications Powerco's have offered reduced connection 3 phases of 20A/phase for near equivalent cost of a 60A single phase supply which offers the convenience of being able to continue to use small 3 phase equipment without the substantially higher cost of a 60A 3 phase supply.

Attached is a pic of the label on a pole mount 50kVA 3 phase transformer installed near our dwelling yesterday replacing the 15 yr old 30kVA transformer that had developed a faults and continually blew its 3A HV fuse last week.
Lucky there was another transformer close by that they could back feed us from until the replacement was installed. 

[max.wwwang]

Thank you tautech. Interesting stuff about the line prices, the 60A current threshold, and the single/three-phase options etc. I never knew this before. Relevantly, at one point I wanted to have three-phase fed to my house for some small reason (which could hardly justify the cost) but didn't get it done (and no idea about the cost).

These transformers (often ABB) around the corners of the streets in residential areas do attract my attention when I come across them.

[max.wwwang]

There are multiple reasons to reference things to earth. One reason is common mode voltage. If the transformer feeding your house didn't have the secondary bonded to earth, it's possible you could get a high common mode voltage due either to capacitive coupling, or worse a leakage or short. In other words live and neutral could both be floating at 7200V or whatever feeds the primary of the distribution transformer. This is as you might imagine, a very bad situation.

That's interesting. But I don't fully (or at all actually) understand common mode voltage.  So this is probably a(nother) dumb question. Since neutral and protection earth (E or PE) are connected (at least at the position of the house switchboard), what's the difference if, after that switchboard, we use only two wires (L/N) and bond the metal case of all consumer appliances to N? This way, touching the metal case will not give you a shock (because it has the same potential as the ground), and a fault of live touching the case will also trip the breaker with the short circuit current. What's wrong with this arrangement? (Note: N is still earthed for the household, but only two wires are used for the sockets all around.)

[max.wwwang]

I still don't know why the scope behaves differently when I flip the polarity of the probe.

[tautech]

I still don't know why the scope behaves differently when I flip the polarity of the probe.

The probe Reference lead is at mains ground potential therefore it shunts any signal to mains ground as srb1954 previously said.

Everyone calls them a ground lead  which indeed they are however as the waveform is referenced against it, it should more correctly be called the Reference lead.
Where you can come unstuck is with isolated channel scopes where each/every Reference lead is isolated from all others and on instruments of this type you can swap leads/probe tips willy nilly just as you might with a DMM.

When the Reference lead on a normal grounded scope pulls the signal down you have the cheapest option of using 2 probes and Maths settings of A-B and just probe a reference point and a POI or get a differential probe which for a few 100 allows you to probe much wherever you need to with 2 properly isolated from mains ground clips.

[max.wwwang]

Thank you tautech. But I don't think I fully understand why the ground lead will shunt the signal in one way but not the other. Can you enlighten me on this?

I was also about to come back and report some new findings.

I said before that when the solenoid was connected, I was unable to detect any voltage across the same two pins both ways. I suspect it was due to the existence of the flyback diode.

Along this line, I've tried with a dummy load of a 30 Ohm (more than double the resistance of the solenoid) pure resistor. This will not bring the diode into function.

With this setup - with the resistance load connected - I am now able to detect the same, but stable, voltage both ways (with of course opposite polarities) with the scope.

Any thoughts on this new information?

[max.wwwang]

Where you can come unstuck is with isolated channel scopes where each/every Reference lead is isolated from all others and on instruments of this type you can swap leads/probe tips willy nilly just as you might with a DMM.

Would love to have one of these isolated scopes. Is this the standard feature of modern models? If so, this may be a good reason to get a newer one!

Alternatively, still with the same old scope, can I do the same thing (nearly) with the scope isolated with an isolation transformer?

[tautech]

Where you can come unstuck is with isolated channel scopes where each/every Reference lead is isolated from all others and on instruments of this type you can swap leads/probe tips willy nilly just as you might with a DMM.

Would love to have one of these isolated scopes. Is this the standard feature of modern models? If so, this may be a good reason to get a newer one!

Alternatively, still with the same old scope, can I do the same thing (nearly) with the scope isolated with an isolation transformer?

No, each channel's Reference lead is still at the same potential. (commoned)

In reality there are few truly isolated channel DSO's on the market as it is an expensive design.
Tek I think TDS models are isolated channel but very small memory for what is expected in a DSO today.
Others and Siglent do some but due to their cost many use a handheld DSO as for at least one channel they provide isolation from mains ground.
However the cost of true isolated DSO's might scare you eg the new Siglent 2ch 100MHz SHS1102X is $2650NZD but it does offer some impressive capabilities:
https://www.eevblog.com/forum/testgear/new-x-model-siglent-handhelds-coming/
https://int.siglent.com/products-overview/shs1000x/

With this setup - with the resistance load connected - I am now able to detect the same, but stable, voltage both ways (with of course opposite polarities) with the scope.

Any thoughts on this new information?

You were lucky.
In the modern DSO we can easily invert any channel to null values from swapping leads.

[srb1954]

Where you can come unstuck is with isolated channel scopes where each/every Reference lead is isolated from all others and on instruments of this type you can swap leads/probe tips willy nilly just as you might with a DMM.

Would love to have one of these isolated scopes. Is this the standard feature of modern models? If so, this may be a good reason to get a newer one!

Alternatively, still with the same old scope, can I do the same thing (nearly) with the scope isolated with an isolation transformer?

It is a very bad idea to power an ordinary scope through an isolation transformer. If the ground lead, or the exposed ground barrel, of a scope probe gets connected, intentionally or accidentally, to a live circuit all the exposed metalwork on the scope also becomes live and a major shock hazard!

For the same reason you should never cut the earth wire on the power lead to the scope.

[IanB]

Would love to have one of these isolated scopes. Is this the standard feature of modern models? If so, this may be a good reason to get a newer one!

Alternatively, still with the same old scope, can I do the same thing (nearly) with the scope isolated with an isolation transformer?

Not with an isolation transformer, but I think you can do the same (nearly) with a differential probe. I think Dave sells one, for example, if they are still in stock. With a differential probe each probe is equivalent and neither is grounded. Therefore you can use them to probe two active points in a circuit without risk of upsetting the circuit under test.

[james_s]

That's interesting. But I don't fully (or at all actually) understand common mode voltage.  So this is probably a(nother) dumb question. Since neutral and protection earth (E or PE) are connected (at least at the position of the house switchboard), what's the difference if, after that switchboard, we use only two wires (L/N) and bond the metal case of all consumer appliances to N? This way, touching the metal case will not give you a shock (because it has the same potential as the ground), and a fault of live touching the case will also trip the breaker with the short circuit current. What's wrong with this arrangement? (Note: N is still earthed for the household, but only two wires are used for the sockets all around.)

Common mode voltage is an offset relative to something, usually ground, that is present on both conductors. For example if a 240V line had a hypothetical 1kV common mode fault, that would mean that neutral was sitting at 1,000V and live at 1,240V.

As I mentioned earlier, neutral is a current carrying conductor. Due to the current passing through it, the resistance of the wire will result in some parts of it being pulled up above ground, possibly by several volts or more. The other issue is that a neutral fault (open circuit) would then result in the cabinet of the appliance floating at mains voltage when no other fault exists. The idea of a protective earth is it is a fallback, no matter what else is going on, any exposed metal should be at earth potential. You're standing on the ground too so you'll be at the same potential and won't get a shock. There is no possible fault condition that is likely to result in protective earth being live.

[TimFox]

That's interesting. But I don't fully (or at all actually) understand common mode voltage.  So this is probably a(nother) dumb question. Since neutral and protection earth (E or PE) are connected (at least at the position of the house switchboard), what's the difference if, after that switchboard, we use only two wires (L/N) and bond the metal case of all consumer appliances to N? This way, touching the metal case will not give you a shock (because it has the same potential as the ground), and a fault of live touching the case will also trip the breaker with the short circuit current. What's wrong with this arrangement? (Note: N is still earthed for the household, but only two wires are used for the sockets all around.)

Common mode voltage is an offset relative to something, usually ground, that is present on both conductors. For example if a 240V line had a hypothetical 1kV common mode fault, that would mean that neutral was sitting at 1,000V and live at 1,240V.

As I mentioned earlier, neutral is a current carrying conductor. Due to the current passing through it, the resistance of the wire will result in some parts of it being pulled up above ground, possibly by several volts or more. The other issue is that a neutral fault (open circuit) would then result in the cabinet of the appliance floating at mains voltage when no other fault exists. The idea of a protective earth is it is a fallback, no matter what else is going on, any exposed metal should be at earth potential. You're standing on the ground too so you'll be at the same potential and won't get a shock. There is no possible fault condition that is likely to result in protective earth being live.

Although a real "neutral" is connected to "ground/protective earth" at some point in the building, due to substantial current flowing in the neutral (white in US) wire between your outlet and the breaker/fuse box where that connection is made, you can have a few volts of AC between the relevant contacts of your outlet when measured with a normal AC voltmeter.

Common-mode definition:  if you have any two voltages A and B, both measured with respect to the same point (for example, the PE contact in your outlet), then by elementary algebra you can define

"D" = difference = differential mode = A - B,   and
"C" = common-mode = (A+B)/2

Thereafter, you can compute A and B from the measured D and C:  A = 2C + D  and  B = 2C - D


Edit:  correction after comment below:  2A = 2C + D  and  2 B = 2C - D.

[Neomys Sapiens]

In reality there are few truly isolated channel DSO's on the market as it is an expensive design.
Tek I think TDS models are isolated channel but very small memory for what is expected in a DSO today.

Not the TDS models, they are very baseline. It's the TPS models. And the Tek portables (THS710/720 and the newest one, whose designation I forgot) There is at least one from Gould too (was also sold by Siemens).

[max.wwwang]

No, each channel's Reference lead is still at the same potential. (commoned)

In reality there are few truly isolated channel DSO's on the market as it is an expensive design.
Tek I think TDS models are isolated channel but very small memory for what is expected in a DSO today.
Others and Siglent do some but due to their cost many use a handheld DSO as for at least one channel they provide isolation from mains ground.
However the cost of true isolated DSO's might scare you eg the new Siglent 2ch 100MHz SHS1102X is $2650NZD but it does offer some impressive capabilities:
https://www.eevblog.com/forum/testgear/new-x-model-siglent-handhelds-coming/
https://int.siglent.com/products-overview/shs1000x/

Oh, that's disappointing.    I thought you said "you can swap leads/probe tips willy nilly just as you might with a DMM". 

I don't understand yet why making DSOs isolated and each channel independent is so difficult.

You were lucky.
In the modern DSO we can easily invert any channel to null values from swapping leads.

What do you mean I was lucky?  Do you mean what I got will not repeat if I do this again?

I understand with a scope that is not isolated, i.e. is earthed, the ground lead will pull the signal down. But the problem is, why is there a stable voltage when I probe both ways, with a dummy resistive load connected?

[max.wwwang]

It is a very bad idea to power an ordinary scope through an isolation transformer. If the ground lead, or the exposed ground barrel, of a scope probe gets connected, intentionally or accidentally, to a live circuit all the exposed metalwork on the scope also becomes live and a major shock hazard!

For the same reason you should never cut the earth wire on the power lead to the scope.

I think I understand what you are saying here. That's the purpose of the earth wire. I only came up with the idea of isolating the scope with a transformer in this discussion - I thought the whole point is, to get more freedom in probing, temporarily disconnect the PE connection, which can be either on the equipment under test or the scope itself.

The risk you are saying is real. But my follow-up question is, if we DO need to use the isolation transformer for the reason of testing (say with the equipment under test), the same risk will exist with the equipment. We can only take extra care to compensate such an increase of risk. The same holds for isolating the scope (which is actually less likely to fail, because the equipment is faulty in the first place). This means no extra increased risk in isolating the scope when using a transformer IS necessary.

Am I missing anything here?

[Edit]
Yes, I am. Refer to reply #41 (answer to question 2) for detail.