Split flap, function of capacitors across stepper coil?

[Sasja]

Hi, im looking at some old splitflaps in order to put them to use again.
So basically they work as a kind of stepper motor, they operate on 24V 150mA current (160 Ohm coil resistance) for about 50ms in one or the other direction to flip to the next character.
As you can see in the pictures, all these devices have capacitors soldered directly across the coil inputs, i've not been able to decode the colors to anything sensible, I measured 4 of those and 3 measured about 4nF and one about 100nF.
I know capacitors are used across dc motors to filter the noise from the brushes sparking and such, but what would be their function in a stepper motor like this without brushes? any ideas?

[Rerouter]

the 50ms setting how is this accomplished?

as only one has the higher value i am almost wondering if it was used to designate start of row, but that doesn't make much sense :/ as unless they are using some fancy multiplexing scheme it should be wiring dependent 

[mrkev]

Well are they really brushless stepper motors? Those would usually have 3 or more coils.

[SeanB]

Those are not capacitors but very old VDR's. Probably silicon carbide. They were used for spark suppression across the coils. The coils are pulsed in sequence to advance to the next character, and have a zero sensor ( usually the blank position) to reset the logic counter in case it got out of sync with the unit. Basically a 50ms pulse on the one direction and after about 10ms a pulse in reverse advances it one character. Pause 50ms between pulse sets to allow the character to finish falling down. First pulse pulls the indexer half way and the second finishes the turn. The VDR's allow a spike for fast turn off of the coils. You might get 100v or so spikes but not enough to kill the drivers.

[Sasja]

thanks! great info!
so, seanB, that would mean i have to see what voltage these vdr's conduct at and make sure my drivers can handle that,
or as a second option, rely on flyback diodes at the driver side (which would limit the turning speed of the flaps)
thanks again for such a quick reply, didnt expect that, really awesome!
 

[SeanB]

Just use 160V rated power transistors and it is likely they will survive. Most of these were powered from +-24V power supplies IIRC, so they only needed 2 transistors per stage with one terminal connected to common. Made a big difference in wiring up a 300 plus digit display saving a lot of expensive transistors at the expense of needing a PNP to the 24v rail and a NPN to the -24V rail, along with level shifting transistors to drive them. You generally had all the digits driven at once as well, so pretty high currents in the power rails for short periods but low overall cnsumption, so they just used a big capacitor and bridge with a smallish transformer, to make it stiff for short pulses.

[Sasja]

I see, now its pretty clear to me how they should be driven safely. Problem is, im starting on this project and someone else already made driverboards with L293 IC H-bridges to be used on a single 24V supply. Now as these L293 are wired in such a way the outputs are never in hi-impedance mode no high voltages should occur and harm the chips. However no clamp diodes (i said flyback before, my mistake) are present, also no internal diodes (that would be the L293D chip). So every time i switch the h-bridge from 24V-0V to 0V-0V 50ms later, the solenoid must be pulling/forcing about 150mA trough a low output of the H-bridge driver. I tried to measure it on my analog scope and find about -1V on that low pin that is sourcing this current to the solenoid. But as there are no diodes, i might be abusing the output pins doing so. The L293 datasheet just states that diodes should be used, not what current a low-impedance output could source reliably. Well the drivers seem to work fine now but... naaaa, i guess im just reluctant to desolder 70 16pin DIPs(L293), buy 70 new (L293D) and solder them in  (adding diodes on the pcb would probably be a even bigger PITA)
but thats probably what needs to be done if dont want the drivers dying one by one in a few weeks or so... ah well.
Thanks again SeanB!

[SeanB]

Drive both low sides on in idle, that will fix the no diode problem by allowing the current to circulate. Slightly higher power in non operation but should not worry the drivers. Just never disable the chips with the enable pin. As the drive is going to be well defined if you want to add diodes you only need one per bridge in any case to clamp the side going low, or one per half bridge as you do not need to clamp to the supply rail in any case.

[Sasja]

yes that was what i was intending. The chips are actually hardwired on the pcb to be always enabled. But, without diodes i would damage the outputs by letting current flow out of a low output right? The datasheet of the chip does show a circuit diagram of the internal output stage and it seems that this current flowing out of an enabled low pin would mean pulling current trough a internal NPN collector that was meant to conduct the other way and such, i dunno im just guessing. You seem to suggest that it could be done safely without diodes, which would be good news. But wouldnt i be abusing these H-bridges then?

[SeanB]

You would be abusing the substrate diodes, and there is a risk of latching up one chip every so often. Might think of putting the diodes into the backplane or the wiring loom as that will be easy to do.

[Sasja]

got it! thank you so much for sharing your knowledge 
hopefully some day I will be able to return all the help i have yet received on this forum!