I need to make a special measuring timer

Sounds to me like a poor stamping fabrication setup. Can't they use a base the parts can sit in that has an outer wall to prevent them from moving off center?

Also, are these parts carbon steel? can you magnetize the table between cycles?

If you really want to do probe sensing, wrap a coil form loosely coupled around the chuck that holds punches. Excite the coil with a ac reference you can monitor. when the punch touches the surface, assuming this is metal we are talking about, it'll create the secondary side of a loosely coupled transformer and change the reference current. This way, you don't need to worry about hard impacts destroying sensors.

Or you can do a E-line reflection detector on the punch.

Jamie

What a description.

But I'm getting the idea that there are two hammers that drop by gravity, lifted by cams that have a sudden edge on the cam allowing the hammers to drop, and that they get out of alignment from each other with use and the cam rotations need to be readjusted periodically. The punches are spring loaded against the work material before the hammers are released to impact them. The problem is that the hammers don't hit simultaneously. You need to work out a method of periodically adjusting _something_ to get the timing true, again.

Is that about it?

Are both cams are on the same drive rod and motor? How exactly are things adjusted (what is adjusted and how?) What are the masses of the hammers?

If you use a piezo and can wire something to them while they are in motion, you might place them at the backside of each hammer. You will get quite a spurt of voltage spikes on impact and they won't wear out located there. A digitizing scope would make it pretty easy to estimate the timing difference, I suppose. Could autocorrelate a digitized sound recording, too, I suppose (the impact sounds of each hammer being similar enough to each other should make the autocorrelation work.) But you wouldn't necessarily know which was first and which second, though you may know that for other reasons. You certainly would know if you got them well matched, though.

Jon

I can't comment on the piezo sensing idea (I think you may just end up with mashed piezos).

But on the DSO side, as long as your DSO actually measures simultaneously you should be fine. That's not a guarantee in a really truly dirt-cheap scope, and I'm not sure how to tell without getting one and finding out.

You might want to stop and think about how much your time and your customer's time costs, then go buy a Rigol or similar inexpensive scope, as opposed to the least expensive dirt cheap thing you can find.

Most piezos are brittle ceramic, which would be a problem here. You will probably want the kind made with plastic film. I don't know models numbers, just heard about this stuff. It's output voltage is lower than the ceramics, so no issues with scope inputs.

You can even use a Windows computer with sound card to do this, since modern sound cards have simultaneous sampling on the A/D converters. Even at the default 48000 Hz sample rate you could resolve to 20.8 microseconds, and most cards (even cheapie built-in laptop chipsets) can sample at 96000 (10.4 usec) or 192000 (5.2 usec) these days.

Sound cards can only handle about +/-2.5 V, so if you are not sure about just how big your signal might get, make a simple limiter: Use a couple of reverse-paralleled LEDs to ground across the input, and feed them with a dropping resistor. See "Sound Card Input Range and Limiter Circuits" at

Best regards,

Bob Masta DAQARTA v7.21 Data AcQuisition And Real-Time Analysis

://dx.com/p/ds-203-3-0-lcd-pocket-mini-oscilloscope-65990

I don't know about the peizo as sensor. I wouldn't worry too much about damaging the scope, but you coould start with some resistance in series with the scope input... maybe start 100k ohm... it depends on how long a cable (how much capacitance.) Or use a x10 scope probe.

Other ways to sense the punch... maybe sound? put some little mic's near each punch? Or could you use a LED and photodiode... attach something to the punch that will break a beam as it moves. Hmm maybe a magnetic reed switch?

George H.

The punch first touches the part and holds it in place. Then it is struck to put the stake in. I need to know when the hammer strikes, not when the punch touches. Eric

Greetings Jon, Your description is pretty close. I haven't actually seen the complete machine, I just make some parts for it. I'm thinking that the piezo could mounted to the underside of a piece of aluminum, the aluminum being struck by the punch. I do need to determine which punch hits first and the time difference between the hits. Eric

Thanks Tim, I'll look at a Rigol. I have never heard of the name. I'm not a real electronics guy, I mostly just fiddle with stuff when I need to make or fix something. Eric

Thanks for advice Bob. I'll check out the plastic piezos. Eric

link:

Greetings George, Since I need to sense impact, not the punch touching the part, a mic might work. If the piezo idea doesn't pan out I'll try the mic. Eric

link:

Perhaps a Strain gauge would solve the measurement problem.

If the signal form is not important you can use the open bench logic sniffer. With this you can easily measure the time between pulses.

Could you use a of microphones to listen to the hammers and a graphing sound recorder software like audacity to analyse the results?

or a dedicated device could be built with a few dollars worth of parts a centre-zero meter on a phase detector on a pair of envelope detectors on a pair of microphones.

I think the real problem is the machine has one hammer too many.

That is kind of relevant if the miss timed punch has already impacted and moved the worked piece. I thought you wanted a way to align it automatically, not probe if for your eyes only?

How are the punches driven? hydraulic, pneumatic, solenoid, rotary cams, explosives, quantum singularity, zpm, splitting quarks, higgs boson?

It would be nice to know a little more? And I am sure casimir polder has nothing to do with this.

Jamie

I like Jasen's (and George's) microphone idea. You can probably do it with Audacity, but I'm more confident recommending Daqarta because I know its real-time trigger capabilities can handle this. (Audacity has an "audio" orientation, while Daqarta is more about measurement and instrumentation.)

You *might* be able to learn everything you need from a single mic. I'd start out with a single mic in any case, moving it around to see what the signal looks like in various positions. The idea is that when the machine is set properly you will get one big "bang", but when one hammer is later than the other you will get a "b-bang".

This will probably be visible with a single mic that is placed equidistant from the two points of impact. You'd adjust the hammer timing until you get one maximal peak... as you change relative timing the peak will get lower and then split into two separate peaks.

You *might* be able to use an ordinary mic, plugged into the ordinary Mic In on the sound card. If you haven't used mics with computers, here are a few caveats: The most common mics these days are electrets, and they are also best for this job because they are small and have fast response (low mass). But they need a DC bias voltage to operate, which is provided by the Mic In connector. The Mic In connection has more gain than Line In, but the sound card limits the frequency response on Mic In (typically 9 kHz instead of 20+ kHz on Line In). There can also be screwy sample rate conversion issues, since Mic In is apparently often done with a separate A/D.

The other issue about Mic In is that it is usually mono only, at least on systems that have a separate Line In. Some of the newer cheapie laptops dispense with Line In connectors and have stereo Mic In, however.

With all that, I'd still recommend trying whatever mic you have on Mic In. If it looks promising, but for whatever reason you feel you really need two mics (and don't have a laptop with stereo Mic In), you will need a separate preamp with connectors for two mics. The preamp will provide the bias for each, and convert to Line In levels. A lot of extra expense and fooling around, however, and I'd bet that you don't need anything more than a simple mic.

If you want to try this with Daqarta, feel free to contact me with any questions using the Contact links on the site. Daqarta has a 30-day / 30-session trial period, which should be more than enough to do this job. If you run short, let me know and I can provide an extension. I'm also glad to provide advice on triggering strategies or anything else. (Note that you can right-click anything in Daqarta and instantly see its Help topic.)

Best regards,

Bob Masta DAQARTA v7.21 Data AcQuisition And Real-Time Analysis

Greetings Jamie, The punches both touch at the same time. The mechanism that hits the punches is what gets out of time. I am not quite sure how the whole thing works, my customer is in a different state. Since I have solved other mechanical engineering problems for this customer he asked if I could help on this one. Somehow the whole thing is adjustable. I make the punches to an exact length and that solves the problem of them making the initial spring loaded contact. The hammer strike is a different mechanism and it gets out of time. Eric

link:

What is an "open bench logic sniffer"? Can you point me to a web site that describes it? Eric

link:

Ever heard of google? Put your question "open bench logic sniffer" in it. You will find:

What it does is, after a start or after a trigger record digital signals. The duration is dependent on the sample rate and buffer size. It can do max 200 Mega samples per second. After the sampling the buffer is read by the controlling program. Then it can be analyzed. In your case setting two cursors on the start of the signals of the strain gauges and read the time difference.

Thanks for the advice John. I have ordered the piezo transducers and will experiment with them when they arrive and let everyone know if I achieve success. Eric