1
\$\begingroup\$

I am trying to use a PCI-Quad04 board to count pulses (from a rotary encoder). In order to test out counting, I have a 1000Hz square-wave signal from a signal generator (BK Precision, 3011B) as input. The signal generator has two outputs - one labeled TTL/CMOS, and one labeled OUTPUT. As I understand - (at least one difference) between these two outputs is that TTL is 50 ohm and OUTPUT is 600 ohm. Signals from both of these outputs look identical on scope (5V p-p). I would expect this - as high impedance scope probe should see 50 and 600 ohm input as pretty similar. (Quad counter inputs should also be high impedance...)

However - in terms of behavior, when signal is coming from OUTPUT, I get perfect # of input counts (1000Hz) from Quad04 counter. However, if I connect to TTL/CMOS output (same amplitude/shape as OUTPUT pulses as shown on scope) of generator - I lose many counts (fluctuating between ~500 - 700Hz, instead of 1000Hz). I am trying to understand where this difference arises from. The reason why this is critical to my application is that I am seeing same behavior as with TTL (fluctuating # of read counts) when I look at pulses from rotary encoder also.

\$\endgroup\$
4
  • \$\begingroup\$ What do the signals look like when you put a load on the 3011B? \$\endgroup\$
    – Voltage Spike
    May 15 at 16:41
  • \$\begingroup\$ I understand - (at least one difference) between these two outputs is that TTL is 50 ohm and OUTPUT is 600 ohm. I checked this and this (the official documents on BK Precision's website were no longer available) but I couldn't see the 600-Ohm output impedance spec anywhere. The output impedance spec is given as 50 Ohms. \$\endgroup\$ May 15 at 16:54
  • \$\begingroup\$ I will have to check on what signals look like with load. \$\endgroup\$
    – RealNovice
    May 15 at 17:09
  • \$\begingroup\$ My understanding that outputs were 50 and 600 ohms was based on link. \$\endgroup\$
    – RealNovice
    May 15 at 17:11

2 Answers 2

2
\$\begingroup\$

According to the PCI-Quad04 manual, input interface uses SN75ALS175 receiver. It has 50mV hysteresis and seems to require input voltage difference to exceed +200 mV and/or -200mV (which doesn't quite seem to be consistent with 50mV hysteresis to me).

This might explain why the straight 50 ohm OUTPUT from the BK precision 3011B function generator works OK, while TTL/CMOS output does not...the right-hand-side BNC OUTPUT gives average voltage of zero, if DC offset is switched off. Its peak amplitude swings between a +ve voltage to a -ve voltage depending on where the amplitude knob is set. This would work well so long as amplitude is set higher than 200 mV.

However, the TTL/CMOS BNC output jack does not swing negatively. It is always a positive voltage that swings between near-zero volts up to about +3V. Its amplitude is high enough, but never swings negative with respect to GND.

Texas instruments show this connection diagram to a function generator for testing. Note that one of the differential inputs is DC-biased at +1.5V. This would work well if the other differential input is driven from a TTL or CMOS source. However, a TTL signal source would object to driving the 50 ohm resistor - it should be removed (open circuit): 75ALS175 data sheet connection diagram

PCI-Quad04 manual shows input connections for two situations (differential versus single-ended). The single-ended situation requires the "+" input to swing to a negative voltage.

Differential:

  • PhaseA, PhaseB and Index (+) inputs at user connector routed to (+) inputs of differential receiver.
  • PhaseA, PhaseB and Index (–) inputs at user connector routed to (–) inputs of differential receiver.

Single-ended:

  • PhaseA, PhaseB and Index (+) inputs at user connector routed to (+) inputs of differential receiver.
  • PhaseA, PhaseB and Index (-) inputs at user connector routed to ground

If you should choose to use the differential jumper, then you should bias the PhaseA (-) input to +1.2V (approximately) for a TTL signal source driving PhaseA (+) input. You could use a clean, bypassed +1.2V DC signal for every (-) input of the PCI-Quad04 connector, if all your input sources are TTL or CMOS.

\$\endgroup\$
1
\$\begingroup\$

As I understand - (at least one difference) between these two outputs is that TTL is 50 ohm and OUTPUT is 600 ohm.

I think you have this reversed. TTL drivers typically can't supply enough current to drive a 50-ohm load, and a function generators are usually designed to supply their analog output to a 50-ohm load.

And this manual also says that the OUTPUT jack is able to produce up to +/- 5 V into a 50 ohm load, which isn't consistent with a 600-ohm output.

Signals from both of these outputs look identical on scope (5V p-p).

According to the manual linked above, the TTL output produces only 3 V p-p. It sounds like you probably have the TTL/CMOS knob pulled out, to produce CMOS outputs. CMOS output amplitude is adjustable between about 4 V and 14.5 V p-p.

if I connect to TTL/CMOS output (same amplitude/shape as OUTPUT pulses as shown on scope) of generator - I lose many counts (fluctuating between ~500 - 700Hz, instead of 1000Hz).

If the CMOS output actually has about 600 ohms impedance and you connect it to a long enough cable, you may be producing reflections on the line that could confuse your event counter.

\$\endgroup\$
1
  • \$\begingroup\$ Something like the reflections which you mention makes a lot of sense. Is there anything I could do to eliminate these reflections? Cable from rotary encoder is fairly long, perhaps ~10 feet... \$\endgroup\$
    – RealNovice
    May 15 at 17:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.