Adding a Timebase Output to an old Scope

[Bazz4CQJ]

Following the recent discussions about sweep generators and wobbulators, I suddenly realised that my old “hand me down” 20MHz Leader oscilloscope has no timebase output .

The service manual is not the best I’ve ever seen, and the reproduction quality is poor.
However, looking at the sweep generator, there is an easy to access test point (TP13), close to the output to the H-amp, which looks to be giving a good saw-tooth output of about 8V (see attached pictures).

Can anyone advise me whether or not I can connect on to that test point and use it to set up timebase output socket, and if so, how best to make the connection? I suppose that ideally, a buffer amp, but any easier option?

Thanks
B

[Oldmadham]

I'm wondering if you could use the trick BWD uses to get its triggering signal from the vert output stage.

They connect a high value resistor (around 220-330k) to the output of that stage, then bring it back to a smaller series resistor, then the triggering circuitry is fed from the junction of the two.

Unfortunately, in their case, if the high value resistor drifts higher over time, the trigger system fails!

I think you may still need a gain stage to get a usable output, but maybe not.

Another trick might be to look for the schematic of a later, or more upmarket Leader which might have such a facility, & copy it, or look at other companies, like Philips or Tektronix.

One of my rules when struggling to understand some weird circuitry, or looking for an obscure part was "See how Philips does it!"
They often had similar circuitry, & had solved the problem, or at least, their version was better illustrated, & they used more obtainable parts.

[Radio Wrangler]

Plan B

If your scope has an X input, then you can just build a little sawtooth generator (dual opamp) that will drive your sweeper, and feed a sample of it to the scope. Adjust the scope shift and gain to position the sweep on the screen.

Some nice scopes allow you to put X in one of the Y channels, Y in the other, so you have loads of control.

David

This way you don't have to mess arounf with a good working scope.

[Bazz4CQJ]

I may have confused the issue; what I'm looking to do is get a timebase output to "drive" a wobbulator.

The thought occurred to me after I made the post that I could just as well take out a signal from the H-amp; there's an easy to access test point on that board with a waveform peaking at ~100V, and it's also very close to the back panel of the chassis where an output socket could be easily placed.

The Haigh wobbulator articles says he used it typically with scopes giving ~25V signals. He has lots of resistance at the input, so it should be a reasonably high impedance load.

Wonder if this makes thing clearer?

B

[Oldmadham]

Originally Posted by Radio Wrangler

Plan B

If your scope has an X input, then you can just build a little sawtooth generator (dual opamp) that will drive your sweeper, and feed a sample of it to the scope. Adjust the scope shift and gain to position the sweep on the screen.

Some nice scopes allow you to put X in one of the Y channels, Y in the other, so you have loads of control.

David

This way you don't have to mess arounf with a good working scope.

From memory, many, (if not most) RF & video sweepers supply a sawtooth signal which can be used to drive the 'scope X input, so Radio Wrangler's suggestion would be a slightly different way of doing the same thing.

Tek 7000 series 'scopes provide an X sweep output, but I never had to use that to drive a sweeper.

[Bazz4CQJ]

OK, got what your saying - build a small external sweeper to drive both the scope and the wobbulator, so avoiding the need to go poking around inside an elderly scope .

Yes, the Leader is two channel and Ch1 will work as an X input, though I've never has reason to use it that way previously.

I guess something like a 555 will do that; what sort of scan rate are we talking about?

Thanks
B

[Oldmadham]

Originally Posted by Bazz4CQJ

OK, got what your saying - build a small external sweeper to drive both the scope and the wobbulator, so avoiding the need to go poking around inside an elderly scope .

Yes, the Leader is two channel and Ch1 will work as an X input, though I've never has reason to use it that way previously.

I guess something like a 555 will do that; what sort of scan rate are we talking about?

Thanks
B

Interestingly, the pix of the front panel of Haigh's wobbulator which turn up when Googling, all show the sweep socket as an output to be connected to the 'scope X input, & all the panel connectors as BNC.

I finally found the original article on scribd.(I didn't sign up).
The pix at the top shows all the connectors as terminal posts, & the sweep one as an input to be fed from the 'scopes X output.

It seems that most people have modified the original design to provide an internal sawtooth generator.

Possibly, Haigh may have done the same in a later article, & this is why all the other pix look like that.

It you Search for "Haigh wobbulator" on this forum, a lot of interesting posts come up, including suggested schematics for the sawtooth generator.

[peter_scott]

The circuit has both + and - 15 volt rails and the TP signal looks to be about 0 to 10 volts. Why not just buffer it with an OP amp connected as a unity gain non-inverting buffer? If you wanted to protect the OP amp output from accidental shorting you could feedback with a series resistor of say 1k and take your actual output from the inverting input.

Peter

[Bazz4CQJ]

I'll take a look for follow-up articles on Haigh, but in the original 2003 paper he is definitely using the scope to provide the timebase.

Peter, yes there's something up to 10V accessible at TP15, but as mentioned, there's up to 100V out of the test point in the H-amp. I think the issue that Wrangler was originally considering that it might be best to leave the covers in place on that old scope and that may well be the sensible way to go (unauthorised tampering... ).

B

[Bazz4CQJ]

Seems that adding a timebase to Haigh is a well-trodden and local path https://vintage-radio.net/forum/showthread.php?t=150034

B

[Wendymott]

The nice thing about an internal sweep oscillator is that you can adjust the sweep length sweep rate and start points easier than using a scope timebase output. I have found my hard copy of the sweep generator I used, but it will take a few mins to convert to a "soft copy".

[Oldmadham]

Originally Posted by Bazz4CQJ

I'll take a look for follow-up articles on Haigh, but in the original 2003 paper he is definitely using the scope to provide the timebase.

Peter, yes there's something up to 10V accessible at TP15, but as mentioned, there's up to 100V out of the test point in the H-amp. I think the issue that Wrangler was originally considering that it might be best to leave the covers in place on that old scope and that may well be the sensible way to go (unauthorised tampering... ).

B

Not to mention that, if you pick up a higher spec 'scope later, your wobbulator will still be useable without any mods.

[Electronpusher0]

Just a quick note.
The thread linked in post 10 concludes with David's redraw of the Armstrong ramp generator
https://vintage-radio.net/forum/show...5&postcount=14
However David has still missed one correction discussed in the thread. R5 should be 10 ohm not 10K (sorry David)

Its a good circuit but I am thinking of replacing the single transistor constant current generator with a current mirror type.

Peter

[David G4EBT]

Thanks for raising that Peter.

At post #7 in the earlier thread Terry, VK5TM stated:

"For the 555 version, R5 is 10R not 10k (10k works because the cmos 555 has such a low current draw compared to the normal 555)".

And at post #8 Dalkemoore2007 stated:

"Arrr That's more logical ~! i was wondering about the 10 k resistor also the 555 is very frequency sensitive to what voltage its run on 10 ohm now that's more like it ".

As I explained, in post #5 of the earlier thread, when I built the generator I used the freehand sketched circuit which Colin Armstrong kindly emailed to me, which I've attached below. You'll see that it clearly shows 10K for R5. As my practical skills and abilities exceed my technical abilities, I didn't understand how the circuits works (or initially, in my case, why it didn't work!) so attached no importance to the value of R5. I'm a bit of a fusspot about freehand sketches, which I never draw for my own use and never re-post anyone else's on the forum, so I re-drew the circuit tidily to make it presentable, (albeit with a few errors, later corrected!).

The second pic below shows the final edit of my own version, in which I simply used the same value of components as the original freehand sketch, (the origin of which I think was from America). As mine didn't initially work, Colin Armstrong said he'd used the CMOS version of the 555 - the LM7555 - so when that was fitted in place of the 555 the generator worked fine. I guess that the reason it didn't work initially may well have been that had R5 been changed to 10R, it would have worked fine, so if anyone wishes to build it and to use a plain 555, I guess that would be the thing to do.

As I built mine into the RB Wobbulator which uses 2 x PP3s in series, I added a 78L12 12V regulator to the circuit and the PCB.

As I have a spare PCB of no use to me but I didn't want to bin, as Barrie (Bazz4CQJ) has shown an interest, I've popped it in the post to him.

[Electronpusher0]

Thanks for the update David.
Actually if R5 is made 10 ohm the circuit will work with either a normal 555 OR the Cmos version.

The circuit is very good. Q1 is a constant current generator that uses 4 series diodes as a voltage reference and R9/R2 set the current.
The constant current charges C1 in a linear fashion. U1a buffers the ramp and provides the output.
The ouput also goes to U1b, a comparator, and when the ramp reaches a threshold set by R3 & R6 the output of U1b goes low.
This triggers the 555 which is configured as a one shot monostable multivibrator.
The output of the 555 on pin 3 goes high for the period determined by R4 & C3, 0.24mS.
This turns on Q2 which discharges C1. When the monostable multibrator times out the output goes low again, Q2 turns off and the cycle repeats.

While the Cmos 555 works with R5 as 10K it will affect the timing of the monostable, I would recommend making R5 10 ohm in all cases.

I am looking to build the Haigh wobbulator hence my interest. I have managed to get equivalent coils and Varicap diode.

The reason I am going to use a current mirror for the current source is temperature stability. The Armstrong design uses 4 diodes as a voltage reference, one will be compensated by the base - emmitter junction of Q1 but that leaves the temperature coefficient of 3 diodes in series, 6mV/degC.
The current mirror is more stable with temperature change provided the two transistors are closely coupled thermally.

The Haigh wobbulator article is also in ETI Vol 27 #13 available from Worldradiohistory
https://worldradiohistory.com/UK/Ele...1998-No-13.pdf


Peter

[Bazz4CQJ]

Just to tie up any loose ends, as David (4EBT) has kindly offered me a PCB for the Armstrong timebase unit, I'm going to proceed with assembling that and put the idea of modifying my old scope in the pending tray.

Looking at the Haigh wobbulator which I have, as stated previously, it has been constructed with just one Toko coil, which has a number which I cannot identify (L132 PC651) and instead of using the suggested varicap diode, it has a 1k533 in it which I assume is a 1N533, which is a zener, so I'll have find out to what frequency it's on and how well the sweep works, but I'm inclined to fit a varicap.

Thanks for all your various inputs.

B

[David G4EBT]

Regarding Peter's helpful comments in post #15 above regarding reducing the value of R5 from 10K down to 10R, as I outlined earlier, the original circuit which I tidied up but with all the same component values, (albeit I added a voltage regulator), specified a plain NE555 and a 10K resistor for R5. It's no wonder that with a 10K resistor for R5 and a plain NE555, mine didn't initially work. There is quite a list of various versions of the 555 at the link below. The different versions of the 555 each have significantly different parameters, notably the current drawn, the minimum supply Voltage and the maximum frequency at which they'll operate.

NE555 current 3000uA, max frequency 0.1MHz Min supply Voltage: 4.5V. (Max 16V).
LM7555 40uA, max frequency 1 MHZ. Min supply Voltage: 2V. (Max 18V).
LMC555 100uA, max frequency 3 MHz. Min supply Voltage 1.5V. (Max 15V).

(In this application the upper frequency limit is of no consequence, but will be in some applications).

https://en.wikipedia.org/wiki/555_ti...as%20of%202013.

So, given that the NE555 needs 3mA and 4.5V, and the LM755 needs only 40uA and 2V, it's not surprising that with a 10k resistor, mine didn't work! With a LM7555 the board worked well with a nice sawtooth output of c. 8.3v P-P and a frequency range of 3 to 117Hz (I set it to 50Hz).

In view of this, I've updated the circuit by changing the value of R5 from 10K to 10R so any version of the 555 should work, and have added a footnote to explain. The updated circuit is attached.

Amazing that the 555 was introduced in 1972 so is almost half a century old, yet still widely used! All the variations are listed here:

https://en.wikipedia.org/wiki/555_timer_IC

Really, what this little escapade reaffirmed to me to me is what I've known since I was a small child - namely, that I'm more of an artisan than a technician. All I needed to have done to pinpoint which the 555 wouldn't work was to have poked the probes of my multi-meter onto pin 8 & Ground to check that the supply voltage on the 555 was at least 4.5V, and no - it wouldn't have been. Not high level stuff is it?

My practical skills have always exceeded my technical skills by a good measure and since the mid 1950s when in my early teens, I have faithfully reproduced every error that was ever published in any magazine project that I've built. As a result, I've left in my trail many nicely produced items which looked pretty but never worked and never would. Time and again down the years, friends have spotted some project or another on my shelves and said 'That looks neat - what does it do?' To which too often I've had to say 'It doesn't do anything - it just sits there gathering dust, but thank you for your kind comments'.

It took me a while to decode Practical Wireless terms such as 'An experimenters' Power Supply'. Loosely translated, 'here's a power supply - it won't work, but if you experiment, you might get it to work. Try not to kill yourself'. Or: 'An Enthusiasts' TRF', meaning: 'If you're super enthusiastic, you might just get this to work before your enthusiasm is utterly and completely dissipated'. Worse still: 'A Beginners' Basic Signal Generator'. So you think 'I'm a beginner, sort of, so I should be able to build this and get it to work'. No - pull it to bits and use the reclaimed components for another doomed project.

But such joy when things do eventually work and how wonderful it is that nowadays, we have internet and excellent forums such as this one to set us on the right track. I've been in it for the long haul - it's 68 years this year since I built my first Hobbies Weekly crystal set (it worked!), and I try to keep in mind Churchill's motto: "Success is the result of going from one failure to another with no loss of enthusiasm".

Sorry - waffling and dribbling again.

[Radio Wrangler]

One standard warning:

Remember that the 555 in its usual bipolar form has a design bug. During the transition of the output stage, the output stage goes unstable producing multiple fast pulses on the output and multiple pulsed attempts to short the power supply. Multiple source suppliers of the parts shared layout works, so just assume they all do it.

CMOS variants of the 555 are quite tame and make good house pets.

It's probably not going to upset a ramp generator with the unexpected output pulses, but if you try triggering a scope during development, it can drive you nuts. However, even in a tolerant application, you still need to be careful about decouplng.

David

[turretslug]

I realise the thread has trended away from the original OP towards a "master sawtooth source" driving both signal generator and 'scope, but I was prompted to take a closer peek into the Hameg 'scope here with its 5V sweep output socket on the front panel- it simply has a Darlington pair (BC239C driving BC237) sniffing the timing RC junction, this pick-off socket being fed from the BC237 emitter via 470 ohms. Surprisingly but encouragingly simple, really. Hameg is a respected, if budget, brand- if they can get away with it, presumably someone seeking a timebase output could use this sort of scheme as a basis. There are obviously offset and amplitude considerations here, but this is very much tu'penny-ha'penny op-amp territory!

[Electronpusher0]

I think you are being too hard on yourself David, we have all faithfully followed magazine projects trusting that they were designed well and would work. It's only when they don't that one starts thinking about how it should work.
If the magazine projects all worked when built we would never learn anything.

Thanks for the circuit update.

Peter