Signal injection

[jrmintz]

Hi All,

What do you use to inject signal into circuits to check for faults? My signal generator has BNC output jacks, could I use a scope probe? I'd hate to screw up a good one, and the 10X ones will act like attenuators if used backwards, right?

Thanks

 

[Kev]

BNC to RCA adapter and then you can use disposable RCA leads to make a variety of other adapter leads.

Add to this some trafos of various ratios and you can get lots of options. Inline pads can make mic levels. Not perfect but fine for signal tracing.
Even better is a battery driven signal generator that is completely earth lifted.

The scope probe is just that ... only for what it was intended.

 

[jrmintz]

Thanks, Kev. That's what I wanted to know.

:thumb:

 

[NewYorkDave]

You can build one of these in a small box with batteries and use it as a handy, low-distortion signal source.

This circuit is actually serving as the master tone generator in our TV studio, and it's been running night and day for over two years now with no problems.

It's powered off 12VDC, but would probably give good service from a 9V battery as well. It would not be hard to implement variable output level if you need it.

The #344 10V, 14mA lamp was used just 'cause I had a bunch of them laying around. Other, similar miniature lamps will work too, with an adjustment of the 2K feedback trimmer.

 

[CJ]

I use a BNC to mini-clip leads. There is a red and black mini clip on the end of the coax cable.
Available anywhere.
cj

 

[jrmintz]

Thanks gents. That oscillator looks easy enough to build.

:thumb: :thumb:

 

[Kev]

To be honest, this is were the magic of DIY truly exists.

The knowledge and a tool box full of adapter leads and simple test devices.
PASSIVE VU METER ... ( calm down Kev :green: )
Not just for your DIY units but for Pro Units.
People just don't test and interface there Pro Units correctly amd then jump into a forum and slag them off.

PA speaker systems.
A simple 1.5 volt Battery on your choice of speaker connector with a switch.

CHECK your woofers are in the same polarity ... if one comes forward ... they ALL should come forward. The number of time I have explained this to friends and they still get it wrong !!!!

"Kev, what did you do to the PA ... it sounds great ? "
:shock:

arrrgghhhh !!!!

 

[PRR]

> What do you use to inject signal into circuits to check for faults?

My finger. :thumb:

 

[soundguy]

I always suspected PRR was an android.

dave

 

[clintrubber]

from NYD:

You can build one of these in a small box with batteries and use it as a handy, low-distortion signal source.

Looks like a nice circuit ! I guess the low-THD could even be improved for next to nothing by some LPF-filtering before (passively) or with U2B (actively). I have a pice of gear that for adjustments requires a THD below 0.02%, so this osc-circuit comes already close. And it could easily be adapted for 1 kHz. Or build a second one for that freq.


BTW, I'm wondering what the THD-contribution to the <0.05% is of the non-inverting use of the opamps (you know, the dist.-mechanism because of unequal source-impedances for JFET-opamps when used non-inv. configuration, as described in various Walt J.-articles).

U2B sees equal impedances for each of its inputs so should be OK and contribute nothing because of its non-lin. input-capacitances - but how about U1A ?

In other words, would scaling-up in impedance of R3, R7 & 'lamp' be worth while or hardly relevant here ?

Thanks,

Peter

 

[NewYorkDave]

Actually, I listed the THD as "< 0.05%" because 0.05% happens to be the noise floor of the old Heathkit THD meter I was using at the time. :wink: In other words, there was no reading above the noise floor of the meter itself. The actual THD could be 0.04% or 0.001% for all I know. Maybe some day I'll take the time to measure again with a more modern instrument.

Reconfiguring U1B as inverting would be a good way to implement variable output level. You just have to be careful to avoid loading the oscillator, U1A--and you'd also have to reference the positive input of U1B to V+/2. As drawn, the pos. input is already biased correctly by the V+/2 offset at the output of U1A.

 

[clintrubber]

from Dave:

Actually, I listed the THD as "< 0.05%" because 0.05% happens to be the noise floor of the old Heathkit THD meter I was using at the time. In other words, there was no reading above the noise floor of the meter itself. The actual THD could be 0.04% or 0.001% for all I know. Maybe some day I'll take the time to measure again with a more modern instrument.

Cool, I'll build one as drawn and mate it with an Audio Precision to see how low the circuit already can get without additional filtering.

Have a good weekend,

Peter

 

[clintrubber]

Just wondering about 400Hz as signal frequency for the circuit above - that freq is something tape-related, right ?
(just curious and myself not having used tape other than 4-trackers so the obvious reason for 400 Hz might have escaped me)

Thanks,

Peter

 

[NewYorkDave]

400Hz used to be the most common test tone frequency in U.S. broadcast, rather than 1kHz.

I still prefer 400Hz because it's a hell of a lot less annoying than 1kHz.

 

[alk509]

Hey Dave, what's a "#344 lamp"? just any old small light bulb or something more exotic?

Thanks!

Al.

 

[clintrubber]

Hey Dave, what's a "#344 lamp"? just any old small light bulb or something more exotic?

It says in the text here that this is a 10V/14mA lamp.

Such a beast is hard to find here though, most lamps do a lot more mA's at that voltage. Couldn't find it at RS & local supplier hasn't anything either. So I'll be trying to scale the feedback-resistors, based on another type of lamp.

I did find the #344 here though:

http://www.specialtyoptical.com/catalog/micro_lamp_344_1776323.htm

...but I guess the overhead for shipping to Europe of a single lamp will multiply that $0.25 beyond everything I guess.

(Anybody in for a '#344 lamp-groupbuy ? :grin: )

Bye,

Peter

 

[clintrubber]

from NYD:

400Hz used to be the most common test tone frequency in U.S. broadcast, rather than 1kHz.

I still prefer 400Hz because it's a hell of a lot less annoying than 1kHz.

I see, thanks. Good reason, I'll do 400Hz.

Bye,

Peter

 

[Guest]

[quote author="PRR"]> What do you use to inject signal into circuits to check for faults?
My finger. [/quote]

Yes, and if injected signal is too much strong, catch ground to the
other hand.

And by timbre of the wrrrrr you hear you can "measure" input
impedance.
If it is deep - high impedance, maybe tube input.
If it is tinny - maybe bipolars.

Prr, good and failure-free measuring equipment.

xvlk

 

[PRR]

> a 10V/14mA lamp. Such a beast is hard to find

Use a 120V or 230V 7-Watt night-light.

It was good enough for Mr Hewlett and Mr Packard.

You can't scale resistances low enough to use common low-volt bulbs, unless you use a loudspeaker-impedance amplifier to drive the feedback loop. Driving a 230V lamp with this several-volt signal won't give a lot of thermal feedback, but 120V will probably work with reasonable impedances (a few hundred to a few K).

There is another way, if you don't need lowest distortion. Use any handy resistances, like 1K and 2K. You need to trim the 2K very exactly, so make it 1K8 fixed and a 500Ω pot. As you fiddle the pot, it will oscillate up to a square-wave, and down to nothing. In between there is a sine, but it won't stabilize. Get it oscillating but on the edge of dropping out. Then put two 6V Zeners in series with a 22K resistor, and put that across the 1K8 resistor. Now when the oscillation is below 6.6V peak, level rises, above 6.6V peak gain drops and it wants to fall. You may have to fiddle a bit, but when this is stable it is VERY stable. The peaks are lightly flat-topped. Increasing the 22K to a higher value reduces the flat-topping, but makes it less tolerant of frequency-network variations.

 

[clintrubber]

from PRR:

> a 10V/14mA lamp. Such a beast is hard to find

Use a 120V or 230V 7-Watt night-light.

This is definitely a US-flavoured circuit ! #344-type lamps, 120V bulbs etc :wink:

It was good enough for Mr Hewlett and Mr Packard.

I got curious by your HP-reference. I've seen the pics of these boxes before, but hadn't read more about what's going on inside. Interesting to learn that the bulb-stabilization originated here.

FWIW:
http://www.hp.com/hpinfo/abouthp/histnfacts/museum/earlyinstruments/0002/index.html
http://www.hp.com/hpinfo/abouthp/histnfacts/museum/earlyinstruments/0002/other/0002patent.pdf

You can't scale resistances low enough to use common low-volt bulbs, unless you use a loudspeaker-impedance amplifier to drive the feedback loop. Driving a 230V lamp with this several-volt signal won't give a lot of thermal feedback, but 120V will probably work with reasonable impedances (a few hundred to a few K).

Closest to the 10V/14mA rating I could find was something like 14V/37mA.
I must say that I don't have a clear picture of the relation of cold vs hot lamp-resistances. The mentioned 120V / 7W does around 2kOhms when it's hot, right ? But with only a few volts on it, how low would it be ?

There is another way, if you don't need lowest distortion. Use any handy resistances, like 1K and 2K. You need to trim the 2K very exactly, so make it 1K8 fixed and a 500Ω pot. As you fiddle the pot, it will oscillate up to a square-wave, and down to nothing. In between there is a sine, but it won't stabilize. Get it oscillating but on the edge of dropping out. Then put two 6V Zeners in series with a 22K resistor, and put that across the 1K8 resistor. Now when the oscillation is below 6.6V peak, level rises, above 6.6V peak gain drops and it wants to fall. You may have to fiddle a bit, but when this is stable it is VERY stable. The peaks are lightly flat-topped. Increasing the 22K to a higher value reduces the flat-topping, but makes it less tolerant of frequency-network variations.

If the lamp-thing doesn't work out OK I can try that one, thanks for the explanation.
I would like the low distortion though - I like it that this can be a nice & handy little battery powered box and still can be used for THD-measurements (looking for about 0.02 or better). I guess that adding filtering can bring the THD down enough though (haven't tried whether that turns out well practically, but since it's about a fixed frequency here it'll be straightforward).

Thanks,

Peter