EEVblog Electronics Community Forum

Electronics => Open Source Hardware => Topic started by: Harvs on January 28, 2017, 05:33:49 am

Title: Poor mans scanning multimeter
Post by: Harvs on January 28, 2017, 05:33:49 am
I had an large excess of DPDT signal relays, so decided to make a scanning card for meters.

The idea is pretty straight forward, it's USB connected as a virtual comm port, you just send it which channel to connect and it disconnects, delays, then connects the required channel.  Then the PC is used to interface with the meter directly and capture the reading.

It has quite a bit of left over GPIO, which my plan is to interface to my HP 3456a GPIB port and trigger in to automate measurements with that meter.

Everything is on Github along with a couple of example Python scripts for using it with a Agilent U124x meter.

I've got most of the parts, boards and relays to build quite a few of these so I may end up offering then for roughly cost of parts if anyone is interested.

Edit: Forgot the github link
https://github.com/harvie256/DMMSwitchBoard
Title: Re: Poor mans scanning multimeter
Post by: fki82 on February 19, 2017, 05:11:15 pm
Looks very nice!
What Relais did you use?
Title: Re: Poor mans scanning multimeter
Post by: Harvs on February 21, 2017, 07:58:34 am
Sorry I couldn't tell you off the top of my head and I'm a few thousand kms from home at the moment.

They're signal relays though that I had left over from another job.
Title: Re: Poor mans scanning multimeter
Post by: daveatol on February 22, 2017, 01:17:21 am
They're PS-5 by the looks of the photos. See http://ultran.ru/sites/default/files/catalog/svetodiody/brend/datasheets/ps.pdf (http://ultran.ru/sites/default/files/catalog/svetodiody/brend/datasheets/ps.pdf)
Title: Re: Poor mans scanning multimeter
Post by: djococaud on February 22, 2017, 09:03:23 am
Very nice !  :-+

I would add polyfuses on each inputs for safety, just in case the micro glitches and select two inputs at once, shorting them...
Not sure if it would greatly affect the readings or not (it would be fine for voltage readings...)

BNC output and GPIB could be a must (but more complicated and not really needed...)
Title: Re: Poor mans scanning multimeter
Post by: voltsandjolts on February 22, 2017, 09:29:19 am
I'd actually thought about doing something similar. Nicely done.
I'll purchase a kit if available.
Title: Re: Poor mans scanning multimeter
Post by: kony on February 22, 2017, 04:12:51 pm
Nice, but no transils on the GPIO headers? Also no bulk decoupling for the relay coils drive section?
Title: Re: Poor mans scanning multimeter
Post by: Harvs on March 03, 2017, 04:12:05 am
Nice, but no transils on the GPIO headers? Also no bulk decoupling for the relay coils drive section?

No... assumption is in most cases the GPIO won't be used, I just broke them out onto a 0.1" PTH pads if needed. There's also no protection on the 3.3V line brought to the headers. Basically like every dev board out there.

Those relays draw a max of ~25mA, with only one on at any one time and they have 0.1uF caps decoupling locally on each one + 10uF ceramic cap at the 5V input.  From memory I believe the USB limit on drive capacitance is 10uF, so in theory I'm already over it.
Title: Re: Poor mans scanning multimeter
Post by: Harvs on March 03, 2017, 04:16:08 am
BNC output and GPIB could be a must (but more complicated and not really needed...)

I use a 0.75" banana to BNC adapter.  They're cheap and work well.

Do you mean GPIB to control something else or GPIB to control it?
Title: Re: Poor mans scanning multimeter
Post by: Kleinstein on March 03, 2017, 08:53:54 am
As is looks, the relays are mono-stable versions. To reduce the power consumption when left on for a longer time, it might be a good idea to use PWM to reduce the current once the relay is engaged. The hold current usually could be something like half the nominal one - thus only 1/4 the power.
This could be important when measuring low voltages.
Title: Re: Poor mans scanning multimeter
Post by: Helix70 on April 03, 2017, 02:42:06 am
If you have any sets of parts left, I would be interested..
Title: Re: Poor mans scanning multimeter
Post by: voltsandjolts on April 04, 2017, 08:12:22 am
...it might be a good idea to use PWM .... This could be important when measuring low voltages.

Are you suggesting this because heat dissipation in the coil would create thermal gradients in the relay, leading to errors from parasitic thermocouples?
Maybe using PWM could create more of a problem with induced noise?

Looking on Digikey, Dual and Single coil latching relays are cheaper than I expected, less than 2GBP each for a decent brand.

Edit:
The MAX4821 relay driver for single-coil latching relays costs more than a relay...although it does drive four of them.

BNC Output: That would be dangerous for higher voltages, could have a live shell.
Title: Re: Poor mans scanning multimeter
Post by: NANDBlog on April 04, 2017, 10:55:02 am
I did similar project for work.
It is missing one or two features, namely:
BNC Trigger output for the multimeter, measurement ready from the multimeter. With this, you can automate it completely.
And catch diodes for the relays, also power supply bypassing.
Title: Re: Poor mans scanning multimeter
Post by: Harvs on April 04, 2017, 11:43:50 am
I did similar project for work.
It is missing one or two features, namely:
BNC Trigger output for the multimeter, measurement ready from the multimeter. With this, you can automate it completely.

That's why I brought out all the GPIO.  I'm triggering with that and reading the GPIB back (well that's the plan at least).  I didn't put the BNCs on the board because I'm working with more than one type of meter that doesn't just use BNC.

Quote
And catch diodes for the relays, also power supply bypassing.

???

There's a RB751S40 catch diode on every relay, 100n cap at every relay & uC + a couple of uF at the regulator.  What is wrong with that?
Title: Re: Poor mans scanning multimeter
Post by: Kilrah on April 04, 2017, 12:08:30 pm
He most likely only quickly looked at the pic where the diodes are hidden, not the schematic as he should have.
Title: Re: Poor mans scanning multimeter
Post by: NANDBlog on April 04, 2017, 12:33:02 pm
He most likely only quickly looked at the pic where the diodes are hidden, not the schematic as he should have.
Not most likely, but defiantly.
I did similar project for work.
It is missing one or two features, namely:
BNC Trigger output for the multimeter, measurement ready from the multimeter. With this, you can automate it completely.

That's why I brought out all the GPIO.  I'm triggering with that and reading the GPIB back (well that's the plan at least).  I didn't put the BNCs on the board because I'm working with more than one type of meter that doesn't just use BNC.

Quote
And catch diodes for the relays, also power supply bypassing.

???

There's a RB751S40 catch diode on every relay, 100n cap at every relay & uC + a couple of uF at the regulator.  What is wrong with that?
Yes, automate it with software. I prefer a fully hardware solution if possible. The multimeter and the devices can be in another room... Well, you sorta can create custom cables for the meter.

Bypassing: you might want to have some bulk capacitor, electrolytic. Also ferrite beads, inductors, for USB, TVS for the power. Relays are nasty.
Title: Re: Poor mans scanning multimeter
Post by: Kleinstein on April 04, 2017, 12:45:13 pm
Catch diode a correct with normal mono-stable relays. They can be tricky with latching relays.

For latching relays one can use a single output and a coupling capacitor - there is no absolute need for special drivers.
Title: Re: Poor mans scanning multimeter
Post by: Harvs on April 04, 2017, 08:47:05 pm
These type of small signal relays are seriously benign.  I think people see a relay and think of your typical power relay that has <<1ohm series resistance and hence is more akin to winding up a inductor.  That's not what these types of relays are.

They have so much series resistance and so little inductance that they are much more akin to switching an LED in, they only draw 25mA max.  E.g. Their effect on the power rail is much much less than something like a single WS1812 LED.
Title: Re: Poor mans scanning multimeter
Post by: Harvs on April 04, 2017, 08:51:29 pm
If you have any sets of parts left, I would be interested..

Sorry I completely missed your post.

I have actually have parts and boards for about 20 of them, but I don't have the time to kit them and send them at the moment.
Title: Re: Poor mans scanning multimeter
Post by: Helix70 on April 05, 2017, 04:10:18 am
If you have any sets of parts left, I would be interested..

Sorry I completely missed your post.

I have actually have parts and boards for about 20 of them, but I don't have the time to kit them and send them at the moment.

Ok, no rush. If you get around to it and remember, I am interested. :)