Touch Probe Amplifier Modifications

[itsme]

Hi there,

I'm going to build a touch probe in the near future and I now have some plans (have a look at the 'Digitizing and Laser Digitizing' section). The probe uses an 'amplifier' so that it uses 12V while the signal to the parallel port is 5V. I have included the schematics for the electronics.

My machine uses a 35VDC power supply and I was wondering how I could modify these electronics to run off 35V, rather than 12V, so that everything is simplified. Here are some of the components that I would potentially use:

MOSFET:
http://uk.farnell.com/jsp/endecaSear...SKU=637488&N=0

Regulator:
http://uk.farnell.com/jsp/endecaSear...KU=4125411&N=0

I am aware that on the schematic it says that the MOSFET will go up in smoke at 16V, however in the data at the above link, it mentions a max voltage of 55V. Any thoughts on this?

One of the other problems is the resistors and capacitors. I would imagine that the sizes of these components on the schematic would be incorrect if I were going to use 35V. Sadly, I have no idea how to work out what these values should be, so any help on this would also be appreciated.

Are there any other suggestions for this? Do my ideas above sound half reasonable (and the components that I've mentioned)?

Thanks
Warren

[Al_The_Man]

As far as I can see you do not need a FET that BIG, look at using a 'Fetlington' 2n7000 (15c here). Also if you are going into a parallel port, I assume you are using a PC, what about using the 5v from the PC supply?
(Did you see this earlier post of mine, actually I picked up the last one from the UK.)http://cnczone.com/forums/showthread...ght=heidenhain
Al.

[gar]

050605-1259 EST USA

itsme:

I like Al's idea of the 2N7000. This is a very useful device. The 2N7000 replaces your Q7.

Here are my additional suggestions.

Noise is always a problem.

Thus, locate the 2N7000 very close to your parallel port input, and the protective components below.

One needs low pass filtering and the threshold point should be made near the middle of the switched voltage range.

For protection of the 2N7000 gate I would put a 100 k 1/4 w resistor from the gate to a point P1. At P1 connect a 0.1 mfd ceramic 20 v capacitor to your GND. Connect the anode of one 1N4148 diode to P1 and the cathode to + 5. This prevents the gate of the 2N7000 from going more positive than 5 V plus the diode drop. Connect a second 1N4148 cathode to P1 and its anode to your GND. This prevents 2N7000 gate from going more negative than GND minus a diode drop. From P1 to P2 use a 10 k 1/4 w resistor. The filter time constant viewed from an input at P2 is 10,000 x .1 microsecond or 1 millisecond.

Hopefully you have -12 v available. Connect a 10 k (power to be determined) resistor from -12 to P2. This increases the time constant to 2 milliseconds. Next connect 1k from P2 to a wire that goes to your probe. ( 050605-1409 edit -- my 1k is a bad choice. Change this to 5.6 k and 1/2 w will be ok. When the switch is closed we will still have positive driving point voltage of 18 v.) The other side of the probe goes to 35 v. Max power to said 10 k is less than 1/4 watt but use a 1/2 w resistor.

Note: the above edit of 1 k to 5.6 k on 050605-1409.
.

[itsme]

Thanks for the reply.

I'm using a laptop on my mill, so getting 5V could be a problem. I think what I might do though, is get a PC case with the power supply and then build all my electronics into that. That way, I'll have a 12V and a 5V supply for things like the probe and fans etc. I still need an enclosure for my electronics and a PC case will give me plenty of space and they are cheap. The only other enclosures I can find in the right size are quite expensive. It'll also mean that I won't need to alter that circuit, which is probably the safest thing (my electronic skills aren't always the best...).

Your probe looks like it was an excellent deal! I looked on eBay and came across a nice Renishaw probe. It was cheap, but it was only the probe head. I then looked for the probe and extensions etc and was soon put off that idea. Those things are costly. I'll just stick to my own brand for now - probably not quite as accurate, but it'll save me a lot of money...

Just as a matter of interest, if I use that MOSFET that I gave the link to, would everything still work OK? Is it OK to use a MOSFET that is too big, besides the extra cost (which really isn't that much more when it's only one...).

My final question is: What sort of resistor would a 470M resistor be (carbon etc)? I just need to know so that I can find it quicker (and easier) on the site where I normally buy this stuff.

Thanks
Warren

[itsme]

Hi Gar,

I was just typing my last message when you posted yours.

I'm really not all that knowledgable when it comes to electronics. I have no idea what a MOSFET does exactly (something like a transistor???), so if it gets too complex, I'm stumped.

As I said, I might just use a PC case. If I do, can I then just use the 5 and 12V supplies from the PC supply and ignore the whole regulator part of the schematic that I posted?

What advantages would a 2N7000 have over the other one? Would it just replace it directly, without any other modifications?

I bet people like me keep you 'electronics guys' entertained/irritated for hours...

Thanks
Warren

[Al_The_Man]

Not only is the 2N7000 MUCH smaller and cheaper, (buy a dozen while you are at it, they always come in handy)! The 2n700 has very high gain due to similarities to a darlington transistor, which is two transistor in one package to create the higher gain. It will replace the bigger fet directly as far as the schematic, the 470m res. can be a 1/4 or 1/8th watt resistor, carbon or glass film.
The original FET is probabally rated higher that the ones in your stepper amplifier
Al.

[gar]

050605-1945 EST USA

itsme:

Some additional reflections.

Change from an FET to a 2N4400 bipolar transistor. Delete my point P1 and the resistor to it. Connect my point P2 to the base of the 2N4400. Connect the emitter to GRD and the collector to the lower end of R3. remove the diode from P2 to +5 because the base to emitter junction serves the same purpose. Make the pull down resistor to -12 a 100 k 1/4 w. Make the input pullup resistor 10 k 1/4 w. Then use +12 v to the probe. The pullup time constant is 1 millisecond, the pulldown constant is 10 milliseconds.

You might put a 1 k 1/4 w in series with the +12 to the probe.

As Al said use carbon composition or metal film. Wire wound would be very expensive and serve no purpose here.

If you built the original circuit with any FET then I would suggest about 47 k to 100 k for R1. It would be better if the lower end of R1 was connected to -12 v.

Note the minimum gate threshold for the IRLZ44N is 1 v. Max and min gate voltage is +16/-16 so this is ok on +/-12 v supply. This will provide very little noise immunity because of the 1 v threshold in the circuit shown. Use Google to search for "international rectifier", then in IR's search box put in IRLZ44N, then pick its data sheet and learn how to read the data sheet.

For the 2N7000 do Google search for Siliconix, then put in 2N7000KL and get its data sheet. Its gate is good to +/- 20 v. If you use this on my circuit with only +/-12 v supply, then previously said protective diodes are not needed. Again in the original circuit that you showed the noise immunity is not good because the 2n7000 minimum threshold voltage is 1 v.

.


.

[gar]

050606-1259 EST USA

On further reflection I like the 2N7000KL as the best choice.

Attached is a DXF file of the circuit. This was drawn for 8.5 x 11 paper.

This circuit must be located close to the TTL input. It requires +/-12 V and about 15 MA from the +12 V side. This has good noise immunity.

Whether my DXF format is compatible or not we will see.

.

[gar]

050607-0843 EST USA

Some comments on my 2N7000 circuit of yesterday.

The threshold point at the 2N7000 gate is about +1 to +3 volts see the Siliconix data sheet.

The input filter to the gate has a 10 millisecond time constant. That is 0.01 x 1,000,000 = 10,000 microseconds. If you want it longer increase either the capacitance or resistance. Keep the lead lengths at the gate short, maybe 1/4 ".

The absolute maximum rating on gate input voltage is +/-20 v. Under normal operation our swing is less than from -12 to +12.

When the probe switch is open the gate has a steady state voltage of aboput -12 v. This is far from the threshold point, and thus provides good noise immunity.

When the probe switch is closed the gate voltage is about 11 volts. Again this gives good noise immunity along with the low pass filter to the gate.

The 1N4733 zener diode clamps the maximum output voltaage to less than 5 v. This is safe for the TTL input.

.