Phase Angle Control Heavy Loads

[Rabbid]

Hi All

I've currently got myself a M5010015F SCR which takes 240V AC and outputs fully rectified DC.

It is layout 1 of the datasheet here http://www.mouser.com/ds/2/657/m_50-1063354.pdf

What's not shown clearly on the datasheet is G1 and G2 both have continuity to each other.

I am attempting to do phase angle control on the output from an Arduino Mega 2560 of this and am currently testing with an old 100W light as a dummy load. I've followed some various guides on how to design this circuit.

Mainly I have the common chips used in the arudino guides.. https://playground.arduino.cc/Main/ACPhaseControl

I've successfully got the input side of the system working using H11AA1 for the zero crossing detector and scoped everything up working lovely. My arduino is outputting with the timer on it's pin just fine for both sides of the AC input.

The problem I'm now having is the output side, the chip recommended in the guides is an MOC3052-M, however the 1.5k resistor following this is causing me troubles. When I took just the live wire and not the neutral I was getting half wave firing working just fine. However when I put a set of diodes on the AC feeding the MOC3052-M I was having issues with the resistor over heating and burning out. Trying larger resistors causes the SCR not to fire at all.

http://docs-europe.electrocomponents.com/webdocs/0d1f/0900766b80d1f182.pdf

From my understanding the larger SCR only needs a pulse of positive voltage on either G1 or G2 to fire then latch until the next zero point. The MOC3052-M chip appears to have a small triac inside it and therefore will latch on connecting G1 or G2 to my MOC3052-M

However I believe my problems may come from the fact it is latching, I am trying to find a coupler which will act as a transistor, but will support a high enough voltage and current rating on the output side.

I've been looking at the data sheet for the MO8204M which may be what I am looking for but I am uncertain as to how to inturpret the datasheet.

http://www.mouser.com/ds/2/149/MOC8204M-196304.pdf

My logic being the arduino pulse will turn on the MOC8204M on for only as long as the arduino is pulsing then turn off again rather than latching, but is this chip suitable for what I need or is it more designed for mains AC on the input?

The board I originally pinched the SCR from used a pair of OEP PT4 pulse transformers attached to an TCA785, this has 2 outputs O1 and O2 based upon whether the AC using crossing to positive or crossing to negative and so the outputs of this chip would fire alternatively through to one of two pulse transformers. When the SCR is off both + and - and G1 and G2 follow the negative DC voltage.

However both transformers output sides are tied to the positive terminal of the SCR output and then through a diode to either G1 or G2 and given that G1 and G2 have continuity my other option would be to replicate a single pulse transformer circuit from the old board and use the arduino to generate a pulse for the for the pulse transformer?

So to summarise, will the MOC8204M do what I want? if not what will?
Would I be better off using a pulse transformer than an optocoupler?

[Rabbid]

These sheets possibly seem clearer,

http://www.farnell.com/datasheets/630968.pdf?_ga=2.203264789.507478389.1511732271-2075370263.1508686603

http://www.farnell.com/datasheets/2050781.pdf?_ga=2.203264789.507478389.1511732271-2075370263.1508686603

[oldway]

Welcome to the forum.

What are the characteristics of the load? Inductive, capacitive or resistive? What voltage and what current?

There are a lot of schematics about phase control of triacs by Arduino, it would have been easier for you to choose another solution: a triac followed by a bridge rectifier.

If you use this power module circuit 1 with an inductive load, you will need to provide a freewheeling diode.

G1 and G2 are not interconnected. The two scr's may have different gate sensitivity....If you trigger both SCR's from one pulse generator, you must provide a 10R resistor in serie with each gate to equalize gate currents.

The description of the module has no information about the trigger sensitivity of the two SCR's.

It is necessary to send a trigger pulse with sufficient current and voltage to ensure a reliable triggering of the thyristor.
For high power scr's, specifications of the pulse where: Uopen circuit : 10V Ishort circuit: 1A peak Rise time: 1µs or less, min lenght: 20µs

The sensitivity of the gate varies with the temperature ... If we use a pulse that is too weak, we risk erratic operation at low temperature.

[Rabbid]

Thank you for your reply,

Eddy brake, 240VAC about 20amps or so per brake and could potentially be 2 or 3 fitted at once. I am basically looking to replace this analog control system for this brake with a digitally controlled system. The current SCR fits nicely in the form factor of the system and would basically just want to design a replacement PCB of the same dimensions and reuse the existing SCR.

This is the F model of the SCR so has an inbuilt flywheeling diode installed inside it.

G1 and G2 do not interconnect on the diagram but with a continuity test on a metre they show as connected and are low resistance to one another.

What about the IC's I've looked at to replace the optotriac with an optotransistor?

Or would it be better to look at the pulse transformers instead?

[oldway]

These informations are important. It is a highly inductive load...!
But a mixed bridge with freewheeling diode is only 1 quadrant, it can't devolve magnetic energy to mains utility.
So, if brake has to be reduced quickly, you should use a 4 scr's bridge.
There is no reason for not to reuse the existing SCR's.

What was the schematic of the power stage ?

In general, I do not recommend changing a circuit if it works properly.

You can measure a low resistance between gate and cathode of each scr, and, as both cathodes are connected, it's normal you measure a low resistance between the two gates.

With highly inductive load, you can't use a short pulse to trigger the scr's....It could result in erratic triggering of the scr's and instability of the control.

You can't trigger these high current scr's directly from an opto-transistor, current is too weak, you will have to provide a pulse amplifier.

This is a project for expert engineer....

[Rabbid]

So really rather than trying to reinvent the wheel, I am best off just implementing it exactly as it was on the old analog PCB which used a pair of pulse transformers, just taking my input from the microcontroller rather than the TCA 785 chip which previously timed the pulse.

Unless you see a better way over the existing setup, this is proven to work reliably.

Attachment 1 shows the main board connection to the SCR.

Attachment 2 and 3 shows a sub board sitting on top of the SCR believe its a snubber network.

[capt bullshot]

Yes, using pulse transformers would be the preferred way to drive these thyristors. The small MOC3052 couplers cannot provide enough gate drive to safely trigger the thyristors. I couldn't find it in the datasheet, but the required gate trigger current can be up to 1A for some 10us, as stated before. So you need either an active driver stage on line potential or use the pulse transformer.

Edit: don't drive the pulse transformers directly from you uC, it's output drivers aren't strong enough. You need some kind of driver stage (like the TCA 785 has built in)

[Rabbid]

Well the pulse transformers used are tied to the + pin of the SCR which will be pulled low to the - when low.

According to the datasheet they have a 200mA peak current and are 1:1 construction.

Electrical specification:
Turns ratio: 1:1
Primary Inductance: 3mH min. @ 1KHz, 0.27V (series equivalent circuit)
Primary DC resistance: 1.1 ohms +/- 10%
Secondary DC resistance: 0.9 ohms +/- 10%
Primary to secondary interwinding capacitance: 20pF
Proof voltage (primary to secondary) 2.8KV rms (peak)
Voltage - Time product: 200V uS
Max. Pulse current: 200mA (peak)
Bandwidth: 3KHz - 1MHz

[Rabbid]

So this 0.2 amps does support what I was seeing as with a 1-2.2k resistor I was seeing the SCR latch on just fine. However with higher resistors it was failing to do so.

But 0.2 amps at 240V would around 50W of power when latching on, would another solution be to swap to a octotransistor which then drives a larger transistor capable of delivering the 200mA pulse.

The current optotiac on the other hand was latching and you can see why the resistor would start to have trouble with the current at higher on times.

[capt bullshot]

So this 0.2 amps does support what I was seeing as with a 1-2.2k resistor I was seeing the SCR latch on just fine. However with higher resistors it was failing to do so.

But 0.2 amps at 240V would around 50W of power when latching on, would another solution be to swap to a octotransistor which then drives a larger transistor capable of delivering the 200mA pulse.

The current optotiac on the other hand was latching and you can see why the resistor would start to have trouble with the current at higher on times.

Usually power dissipation in the gate resistor isn't a problem when triggering a SCR with an optotriac. The MOC should be able to drive 200mA. Latching behaviour is intended here, since the SCR will latch on also which in turn reduces the voltage drop across the resistor to near zero. So it's only pulse loading to be considered for the resistor.
You'll have to connect the optotriac from SCR anode to gate, and have to take care that the optotriac doesn't fire if the SCR is in reverse polarity (when it is supposed to block current). If you fire the optotriac with SCR in reverse, gate current will flow through the triac (which works both polarities), but the SCR won't trigger and then you have a bad situation that causes high losses in the resistor and may harm or cause misbehaviour the SCR. Easiest way to ensure correct triggering would be to put a suitable diode in series to the optotriac so it works only one direction.

Edit: You'll have to use two optotriac couplers, one for each SCR, since the anodes of the thyristors are on different potential, and only one SCR is to be fired at one time (one for positive half wave of line voltage, the other for the negative half wave).

Edit 2: The original pulse transformers use a common pin, because they need to be connected from cathode to gate, that's different from the optotriacs.

[Rabbid]

So the TCA chip had 2 outputs one for going positive and one for going negative, at the minute I have a simple zero crossing.

Should I switch my zero detection to have same functionality so I can fire G1 or G2 individually, or should I simply just put diodes on the feeds to the MOCs?

I do fail to understand why you would want the MOC to latch on if at the end of the day its only job is really to provide a pulse and why the optotransistor isn't more commonly used?

[jbb]

G1 and G2 do not interconnect on the diagram but with a continuity test on a metre they show as connected and are low resistance to one another.

It's possible that your module is damaged, which won't help.  What's the resistance, exactly? In Ohms?

I would expect the gate (G) to cathode (K) to look somewhat like a diode, not a resistor.  I suggest you check using several resistance ranges on your multimeter (a resistor will give consistent readings, a 'diode' will give different measurements at different ranges). And swap the polarity to see what happens.

Well the pulse transformers used are tied to the + pin of the SCR which will be pulled low to the - when low.

According to the datasheet they have a 200mA peak current and are 1:1 construction.

This matches the TCA785 (250mA) drivers.

Also, I suggest you use the proper terms for the thyristors or we'll all get confused and give you bad advice: anode (A), cathode (K), gate (G).

[Rabbid]

I would hope it wasn't damaged as it was pulled out of a known working dyno.

I've re-rigged it up with two MOC's and two 2.2k resistors but I am only seeing triggering when live is positive..

What I've done so far is

Live to ->| to MOC to 2.2kR to G1

Neutral to ->| to MOC to 2.2kR to G2

It appears upon scoping it the negative live is still not triggering. I would have thought that this should work as is G1's anode not AC1 (Live when positive) and G2's anode not AC2 (Neutral when live is negative).

I've also had no issues with burning out resistors now I understand that G1 and G2 are in fact separate. I've used a 2W 2.2K and infrared shows its hitting no more than 10C above ambient when commanding 100% power. Certainly any weirdness or problems were my lack of understanding of this point.

Attached is a scope of it, Cyan connected to +, Pink connected to - and dark blue connected to the arduino firing signal.

[Rabbid]

Black lead - Red lead probing:
High to all terminals

Red lead - Black lead probing:
G1 520ohm
G2 520ohm
+ 475ohms
AC1 450ohms
AC2 450ohms

Black lead + Red lead probing
G1 39ohms
G2 39ohms
- 462 ohms
AC1 High
AC2 High

Red lead + Black lead probing
G1 39ohms
G2 39ohms
- High
AC1 High
AC2 High

AC1/2 to anything is High either way around

G1 to G2 is 70ohms

[capt bullshot]

An SCR's gate to Cathode looks like a low value resistor if probed with a DMM. So your reading of 39 Ohms / 70 Ohm is OK.

Can't exactly tell from the scope picture what is working and what not, looks a bit non-OK.

You'll run into problems with the high value gate resistors when the firing angle is near 0° or 180°, when the voltage difference over the SCR is low. Then there will not be enough gate current to fire the SCR. You could lower the resistors, that in turn brings you higher peak currents near 90° firing angle, they might be too high for the MOC coupler (check data sheet for maximum short pulse current and calculate minimum resistor for this current a maximum voltage). This might result in the 220 Ohms ballpark. May still be unreliable at the range ends of firing angle.
This is the reason why firing the SCRs using pulse transformes is the preferred method. Next valid method would be to provide an isolated 8...12V supply and use some transistor couplers with a driver stage. That way a sufficient gate current is maintained over the whole firing angle range.

Edit: Fire the correct (forward) SCR only. Don't fire the reverse SCR - say copy the TCA 785's firing scheme. Firing both half waves is for triacs only, not for SCRs.

[Rabbid]

The issue with the scope is that it shows a half bridge rectification, where as I'd expect the negative to be at -240V and positive at 0V on the second half cycle with full rectification.

What I see is both + and - showing -240V on the second half cycle so G2 isn't currently firing. G1 disconnected I get no power output. G2 disconnected I get the same power output.

I think this actual MOC is recommended for upto 60mAmp so I may have to get a higher current transistor to safely drive the gates in the final design.

[Rabbid]

What I currently see if I remove the ardunio signal vs what I expect to see if I was obtaining full wave rectification.

Remember the scopes are reverence to earth at 0V, Cyan shows positive output and pink negative output.

[Rabbid]

Issue resolved, my connect of neutral via a diode to a second MOC was right, however the jumper wire on the output to the second gate was faulty.

Basic diagnostics first... 

So yes basically all my problems were I tried to rectify then control only from G1 assuming they were linked. Incorrect.

I needed to supply live though a diode / resistor / MOC for G1 and then neutral through the same circuit duplicated to G2.

My next thoughts are while this appears to be working I believe the current peaks through the MOC maybe too high, especially if I increase the resistor so now I wish to add a transistor to the output of the MOC which will be driven from after the diode and another resistor to be able to handle around 250V 1 amp pulse peak.

[David Hess]

What's not shown clearly on the datasheet is G1 and G2 both have continuity to each other.

G1 and G2 do not interconnect on the diagram but with a continuity test on a metre they show as connected and are low resistance to one another.

Usually the metalization connecting to an SCR's gate deliberately covers part of the cathode acting as a distributed gate-to-cathode resistor to remove charge from the gate-to-cathode junction improving turn off capability.  So when you measure between the SCR's gate and cathode, it looks like a relatively low valued resistor.  In this case since the SCR cathodes are connected together, a gate to gate measurement reveals twice this resistance.

Sensitive gate SCRs lack overlapping gate metalization compromising their turn off capability but making them more sensitive.

[vw1320]

Sent you a PM Rabbid.