Advice request: optical solid state relays

for all subjects/topics not covered by the other forum categories
Post Reply
User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Advice request: optical solid state relays

Post by jonb » Tue Oct 08, 2019 12:39 pm

Hi

This is a bit of a specific request. I have a classic car with a faulty electromechanical clock, which I want to repair. The clock has a solenoid which is used to wind the mechanism. The clock then runs for about two minutes, at which point a pair of contacts close, triggering the solenoid which in turn winds the clock back up and opens the contacts.

This approach has flaws that lead to reliability problems:
  • The trigger contacts are closed very slowly, being connected to the clock winder. This leads to a very slow touchdown and I've discovered that some current can flow before the contacts are fully closed, enough to partially trigger the solenoid but not enough to wind the clock properly and reopen the contacts. This leads to an unpleasant buzzing noise.
  • If the car's battery gets run down (very common occurrence on a car that is used infrequently), the contacts can close but there won't be enough energy to drive the solenoid to re-open them. Current then flows through the coil and the whole thing heats up. This can cause catastrophic damage to the solenoid or minute hand gear, which is plastic.
So, in order to address the first problem, I thought I might wire in a miniature opto isolated solid state relay.

IMG_0014.jpg

The voltage divider is there to provide a low voltage to drive the LED inside the relay. I probably have the LED itself upside down, please ignore. :oops:

The relay's datasheet is shown in the pictures of this eBay listing https://www.ebay.co.uk/itm/302831495739.

One thing not shown is a 100 ohm resistor across the solenoid coil which I suppose is acting as a current limiting device.

So the question is, will this work? Or would a different circuit that gives a timed pulse to guarantee the solenoid is energised for long enough to throw the contacts open be better? I would just clean the contacts but I did that when I had the clock apart for its initial repair.

To address the second point, the attached document outlines a control circuit that protects the clock in the event of shorts or other problems. It says the 16uF capacitor has enough energy to run the clock but I think it can actuate the solenoid only once per charge. Perhaps it is acting as a reservoir that is capable of charging when the battery is low, then releasing the charge as a slug of current that does actuate the solenoid..?

The other thing is the spark arrester circuit. Is it worth fitting this instead of, or as well as the relay? Space is very tight inside the clock but I think I could get both in there..
Attachments
Kienzle Clock Repair2.doc
Clock repair doc
(36.5 KiB) Downloaded 6 times

User avatar
daveejhitchins
Posts: 5114
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham
Contact:

Re: Advice request: optical solid state relays

Post by daveejhitchins » Tue Oct 08, 2019 5:20 pm

Could you not use a micro switch? they have a 'snap' action that would sort your problem out!

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ARA III, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Wed Oct 09, 2019 11:38 am

Hi Jon

I’m fairly certain that your circuit will not do what you want. Also R2 (27Ω) as drawn will only act as a heater and power waster :shock:

If I understand you, what you want is a circuit that
  • has a delay before it operates the output
  • has hysteresis like properties so that once on, it continues to run rather than jittering.
  • only operates for a defined amount of time before turning off
  • has a suitable output to drive the solenoid
  • is small
Is this correct?

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Wed Oct 09, 2019 2:10 pm

Well..

> has a delay before it operates the output

No, it needs to trigger the solenoid when the contacts close.

> has hysteresis like properties so that once on, it continues to run rather than jittering.

Yes, but only to trigger the solenoid with a pulse that is wide / strong enough to pull the winding mechanism.

> only operates for a defined amount of time before turning off

Yes, that's what I mean by pulse.

> has a suitable output to drive the solenoid
> is small

Right on both counts. Space is tight:

InkedIMG_6041_LI.jpg
Red: Contacts. Yellow: Winding wheel. Cyan: PCB

I chose the relay mentioned because it really is tiny - about 5x5mm.

Since posting I had a closer look at it and cleaned the contacts again. I found an adjuster that sets the contact position, and have been able to align them properly so that they make better.. err.. "contact". The clock is now running and I've had it going for about a day. No buzzing yet but I can imagine that once the contacts are dirtied again it'll buzz. As it is difficult to get to once installed, I am trying to improve its reliability beforehand.

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Thu Oct 10, 2019 8:58 am

Darn it. I had the clock running nicely and left it to soak test overnight. For some reason the contacts got stuck closed and the solenoid has burned out. It was running fine all day yesterday and I was setting its speed (what cock people call "regulating" it). I have spares but I need some help to work out a proper protection circuit..

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Thu Oct 10, 2019 12:37 pm

I was thinking of using a simple timing circuit that then controls a transistor. Once triggered, the timer would switch on the transistor for a defined time period, producing an output pulse of current to operate the solenoid.

How long is the pulse to the solenoid? Are we talking hundreds of milliseconds, a couple of seconds or what?
And how much current does the coil of the solenoid require?

A resistor feeding a large electrolytic capacitor can provide protection, but the electrolytic capacitor may be fairly large and given the space limitations...

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Thu Oct 10, 2019 1:00 pm

I'd have to get the scope out for that.. :)

I measure a 1.5uS pulse across the solenoid when the contacts close. It looks like this:

DS1Z_QuickPrint2.png

I assume the pulse width is commensurate with the time it takes for the solenoid to energise and the contacts to open. The action of opening the contacts throws the winder wheel round by about 200 degrees, rather energetically. One of the contacts is mounted on it, and it's grounded. So one side of the solenoid is connected to +12v permanently and the other to the contact arm.

winder mech.JPG
Winder mechanism

The winder wheel is spring loaded so as to return back to the "contacts closed" position. When the contacts close, the wheel is thrown clockwise and the spring loaded pawl engages on the teeth of a serrated gear wheel mounted directly above the winder wheel on the same rotational axis. This keeps the winder wheel in the open position. The winder wheel spring applies torque to the serrated gear wheel via the pawl. The serrated gear wheel drives the clock mechanism, so as time passes it rotates, this in turn allows the winder wheel to rotate (counter clockwise). After about 2 minutes, the contacts close and the solenoid fires up again. This pushes the contact arm violently into the contact on the winder wheel and inertia causes the wheel to rotate clockwise again, thus winding the clock. This only works, I think, because the wheel has some weight to it.

In order to solve the issues plaguing this mechanism, we need to a) reduce the current flowing through the contacts, b) implement a spark arrester for the contacts, c) ensure that the solenoid never gets more than a 1.5uS pulse every time the contacts close and d) ensure there is always enough power to throw the contacts open (so, shut down the clock if the battery runs too low).

I think the document covers b) and d).
Last edited by jonb on Thu Oct 10, 2019 1:58 pm, edited 1 time in total.

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Thu Oct 10, 2019 1:08 pm

Use a 100Ω resistor in each leg (+ & -) between the solenoid and the ‘scope, with a 1N4148 diode wired anode to + and cathode to - (so that normally it will not conduct). That will provide a good level of protection for the ‘scope.

If you know the resistance of the coils of the solenoid, that can give an indication of its current requirements.

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Thu Oct 10, 2019 2:07 pm

Mark, I updated the previous post with a description of the winding mechanism.

I measured the coil and it is 12.4 ohms but remember there's a 100 Ohm resistor wired across it. So the solenoid on its own should measure 14.15 Ohms.

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Thu Oct 10, 2019 8:12 pm

jonb wrote:
Thu Oct 10, 2019 1:00 pm
I measure a 1.5uS pulse across the solenoid when the contacts close.
Wow, that is an amazingly short pulse from an electro-mechanical system :shock:

I have to ask, are you sure about this?

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Thu Oct 10, 2019 8:24 pm

The scope trace isn’t lying.

That was measured across the solenoid connections using single shot. It triggered when the solenoid fired. Have I misinterpreted it?

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Fri Oct 11, 2019 8:43 am

Mark, I'm going to try building the protection circuit that is shown in the document attached to the first post. The only thing I'm not sure about is the capacitor it uses to store energy to kick the solenoid when the battery runs low. I assume what the circuit does is to charge the cap while the clock is running (contacts open) and then release the energy at contact closure; the 470R + 68R resistors providing low current for charging the cap, but not enough to burn out the coil in the event of the mechanism sticking. The capacitor, described as "large" in the doc, is "16mF" - do I assume it means uF and if so, isn't that a bit small to provide a 12v pulse at what I have just discovered is about 1 amp?

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Fri Oct 11, 2019 11:53 am

I would NOT assume 16mF means 16uF, as I don’t think a 16uF capacitor would store enough energy. In fact,
I’m certain that 16uF would be too low.
I was perplexed when I saw that value in the diagram. As it is unusual to use mF and capacitors with values of 16000uF are available, but are large and expensive.

You could try experimenting with a 1000uF 16V capacitor. If that is not enough try two or more in parallel.

I had a play with a NE555 timer circuit last night in a SPICE application, but this wouldn’t produce a pulse anywhere near as short as 1.5us. I’m thinking this is a limitation of the simulation. But it could also be that the simulation is reflecting propagation delays in the chip. In any event 1.5us is a very short pulse. Hence why I was surprised.

I can show you a schematic of the circuit if you want. It can produce a pulse of 500us (including turn off time for a MOSFET output transistor) or longer.

Other designs are possible.

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Fri Oct 11, 2019 12:44 pm

The guy who designed the protection circuit posted another document (attached).

I think this is a bit better, his waveforms show a 300v EMF spike after the solenoid is triggered.

Maybe my scope setting is all wrong!

He also clarified that it’s 16 millifarads or 160000uF. Blimey. I’ll experiment a bit with smaller values..
Attachments
Kienzle%20Clock%20Waveforms.doc
(39 KiB) Downloaded 6 times

User avatar
1024MAK
Posts: 9236
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Advice request: optical solid state relays

Post by 1024MAK » Fri Oct 11, 2019 1:30 pm

So from the ‘scope display images, the pulses look to be around 15ms to 25ms long.

The current waveform is a good demonstration of how the current through an inductor ramps up from zero up to a value limited by the resistance of the wire.

I’m not sure why the author decided to include a Zener diode. A simple silicon diode (a fast 1A diode) is all that is normally used.

Mark

User avatar
jonb
Posts: 2561
Joined: Sat May 21, 2011 12:42 pm
Location: South Coast of England
Contact:

Re: Advice request: optical solid state relays

Post by jonb » Sat Oct 12, 2019 11:22 am

This is the protection circuit with 4x4700 16v banked capacitors.
KCProtector.jpg
I had a discussion with the author and he thinks 10,000uf might do it. He explained that the circuit works as I thought - caps charge on the low power let through by the 470R +68R resistors, then discharge when the clock contacts close. If the clock shorts or jams, the low current through those resistors will prevent it overheating. Neat.

Post Reply