UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Test Gear and Workshop Equipment (https://www.vintage-radio.net/forum/forumdisplay.php?f=36)
-   -   Oscilloscope bandwidth (https://www.vintage-radio.net/forum/showthread.php?t=147360)

radiograham 12th Jun 2018 11:36 pm

Oscilloscope bandwidth
 
Does anyone know if there is a simple add on, or mod, to increase the bandwidth of an oscilloscope. I have a GEC miniscope and want to use it at higher frequencies than it is capable of as it is. Graham.

MrBungle 12th Jun 2018 11:38 pm

Re: Oscilloscope bandwidth
 
Can’t really do that. If it’s an RF signal you can use a demodulator probe or downmix what you are interested in to the bandwidth of the oscilloscope but that’s about it.

turretslug 13th Jun 2018 12:28 am

Re: Oscilloscope bandwidth
 
I suspect that the bandwidth is limited primarily by the output (Y plate driver) stage, i.e. the one with the combination of largest amplitude swing and highest circuit capacitance.

Craig Sawyers 13th Jun 2018 6:42 am

Re: Oscilloscope bandwidth
 
I agree. It is not really possible to increase the bandwidth of an oscilloscope. That is why Tektronix (for example) made a wide range of oscilloscopes with different bandwidths at different price points. At least from the 50's to the 90's.

Since the neat little scope you have only goes from 50Hz to 300kHz, you can do a lot better by spending not a ridiculous sum for a larger scope that goes much higher than that.

Craig

Radio Wrangler 13th Jun 2018 7:34 am

Re: Oscilloscope bandwidth
 
There is such a thing.

Bradley and a few other firms made sampling scope adaptor boxes to be connected ahead of low bandwidth scopes (the scope used as an X-Y display) and the frequency range displayed could go to several hundred MHz.

There were several disadvantages:

Low impedance inputs, usually 50 Ohms.
Couldn't handle many volts
Worked only on repetitive signals
Expensive new.
Fast speed timebases only.
Impractical to keep connecting and removing the adaptor when slower timebase speed was needed.

Nobody really wants these things now, so they get broken up for the knobs, case, PSU.

So, even though they exist, it's a lot easier to just get a higher performance scope. Scopes are herd animals and it's cruel to keep one alone.

David

MrBungle 13th Jun 2018 8:23 am

Re: Oscilloscope bandwidth
 
Good point with sampling.

Also agree about herd animals. I’ve got five at the moment and only one works :)

barrymagrec 13th Jun 2018 8:39 am

Re: Oscilloscope bandwidth
 
With you there - I have seven - and only one works - I desperately need to fix one of the broken ones before I have no working scope at all.

peter_scott 13th Jun 2018 8:39 am

Re: Oscilloscope bandwidth
 
You could plot the present response and design an equaliser to balance out the roll-off. Clearly, if the equaliser was passive you stand to lose a lot of sensitivity but you could make it active although you may then risk over driving the scope's input circuitry.

Peter

GMB 13th Jun 2018 9:10 am

Re: Oscilloscope bandwidth
 
If a scope gives direct access to its plates then it should be possible to get a bit more bandwidth using an external amplifier built using modern parts.

Also, I vaguely remember articles in PW and the like for mods to scopes to up their bandwidth, implying that the built-in amplifiers are not as good as they could be, i.e. the Y amp is usually the limiting factor.

mark_in_manc 13th Jun 2018 9:13 am

Re: Oscilloscope bandwidth
 
I have two Thurlby 'Digital Storage Adpators' DSA524, one of which would do what you want to do as well as considerably more (including dual beam, if you can see them on a screen that small :) ). They're also quite small if space is an issue and you want to keep your existing scope. More about them in this thread, and a user manual at post #6

https://www.vintage-radio.net/forum/...d.php?t=142128

I see you're in Manchester - I'm in Whalley Range. PM me if you'd like to bring your scope over and try one out. If it was any use to you, it would be cheap.

cheers
Mark

radiograham 13th Jun 2018 11:37 pm

Re: Oscilloscope bandwidth
 
Quote:

Originally Posted by Craig Sawyers (Post 1051414)
I agree. It is not really possible to increase the bandwidth of an oscilloscope. That is why Tektronix (for example) made a wide range of oscilloscopes with different bandwidths at different price points. At least from the 50's to the 90's.

Since the neat little scope you have only goes from 50Hz to 300kHz, you can do a lot better by spending not a ridiculous sum for a larger scope that goes much higher than that.

Craig

Am I correct in thinking then that it is the design of the Y amp that determines the bandwidth of a scope,So on a more expensive model this is what gives it the higher bandwidth.I was trying to check the output of my Advance signal genny using my scope but this only works up about 200/250kcs,Where as the genny goes up to 100mcs. Graham.

Cobaltblue 13th Jun 2018 11:47 pm

Re: Oscilloscope bandwidth
 
The Mini Scope due to space has a very simple and primitive Y amp

The tube itself has limitations as well, It was designed as a simple serviscope.

You can improve it but with the available space you would need to add Silicon.

Scopes with better bandwith are avaiable for less than £50

So you need to decide original or modify :thumbsup:

Cheers

Mike T

Bazz4CQJ 14th Jun 2018 12:20 am

Re: Oscilloscope bandwidth
 
1 Attachment(s)
Attached is a circuit from the 2008 ARRL Handbook intended to act as an adapter (convertor?) to enable a scope with a limit of 5MHz to work at 30MHz.

I've got as far as buying the mixer :).

B

Radio Wrangler 14th Jun 2018 4:32 am

Re: Oscilloscope bandwidth
 
A scope isn't going to have any better bandwidth than the Y amplifier, but the CRT also has a high frequency roll-off.

The speed of the electrons passing a pair of deflection plates is high, but still finite. Each electron spends a little time under the influence of the Y plates. During this time, the Y signal changes. The amount of deflection achieved is the integral of the signal over the time the electron was between the plates.

If the Y signal frequency is high, some of the signal starts to cancel what its earlier self did. Eventually when the time in the deflectors equals a full cycle, we get complete cancellation and we see no signal on the screen. Go higher still and the signal grows again, but never as big as before. The frequency respponse is a (sin(x))/x set of bounces.

This is the same mechanism as sets the rolloff of a magnetic tape replay head due to the finite gap and the recorded wavelength.

If we fix the problem by speeding the electrons up with more accelerating voltage, or if we cut the length of the plates, thebandwidth improves but as the electrons are being deflected for less time and we get less deflection at all frequencies. The CRT deflection plates become less sensitive and then we need more voltage swing from the Y amplifier and it gets harder to reach the wanted bandwidth. So faster scopes run hotter.

You just can't win.

So you cheat. Divide long Y plates into several pairs of short 'platelets'. Connect each pair to subsequent tappings on a balanced pair of delay lines. Put a terminating resistor on the end of the line to prevent echoes. Your CRT now has the deflection sensitivity of the length of all the plates added together. It has the bandwidth of the length of one platelet. Neat?

You can do a similar trick by distributing several valves down transmission lines rather than having them in simple parallel. You can build a much faster Y amp this way and now you know why some Tek scopes had a hundred valves in them.

Scopes are fascinating, and the programs are better than on TV.

David

Craig Sawyers 14th Jun 2018 7:18 am

Re: Oscilloscope bandwidth
 
Quote:

Originally Posted by Radio Wrangler (Post 1051630)
You can do a similar trick by distributing several valves down transmission lines rather than having them in simple parallel. You can build a much faster Y amp this way and now you know why some Tek scopes had a hundred valves in them.

Scopes are fascinating, and the programs are better than on TV.

David

Indeed. The 545A has a fully differential Y amp, which includes 6 DK6's on + and 6 DK6's on - in a distributed transmission line structure. They feed into a delay line to slow down the vertical signal to wait for the timebase to get going. That itself has fifty adjustable capacitors (and two inductors) that have to be tweaked in sequence to optimise the rise time.

The fastest plug in amplifier of the day was the Type L, which pushed the 545A to its limit bandwidth of 30MHz. The fastest of the valved scopes (well actually a hybrid with some transistors) was the 585. Tek tried hard to break the 100MHz barrier, and failed - the 585 makes 85MHz.

But sampling plugins like the very easy to use 1S1 takes any of those plug-on scopes to 1GHz.

The distributed amp concept came from Howard Vollum (Tek's founder) and David Hewlett (HP's co-founder) having dinner together and sketching ideas on a table napkin. They were good friends in competition.

Craig

Radio Wrangler 14th Jun 2018 8:43 am

Re: Oscilloscope bandwidth
 
The big paper that set the distributed amplifier going, showing how to design one and also a trick with tapered transmission lines was by Ginzton, Hewlett, Jasberg and Noe. I have a paper copy somewhere. There was a later book "Distributed Amplification" from the Stanford University Microwave Laboratory.

In the paper, they don't claim originality for the idea but trace it back to Britain, and if I remember correctly someone by the name of Percival. So it looks like distributed amplifiers should be added to the list of hovercraft, television, radar, penicillin, hovercraft, steam engine and hovercraft which is often trotted out in documentaries. (They never mention shrapnel, napalm and concentration camps which were also British inventions. We can be nasty devils when roused.)

Ginzton was a Ukrainian whose family fled the Russian revolution and eventually wound up on Fred Terman's postgrad course at Stanford alongside Bill Hewlett. He went on to join the Varian brothers.

Bill Hewlett is well enough known

Noe was one of the early microwave folk at HP and a well known name in the firm.

Jasberg, I don't know anything about.

Howard Vollum came into things a bit later. He'd invented the triggered timebase, which was to revolutionise oscilloscopes. He wanted to sell his invention to Bill and Dave. They liked it and understood what it was worth, but the HP company was founded and growing at this time, and it had its hands full with government orders for microwave instruments. So they advised Howard to start his own firm up in a garage as they had. He did. He called it Tektronix. Bill Dave and Howard remained friends. HP and Tektronix had a formal agreement giving each firm free licence to the other's patents. There was rivalry between the two firms, but it was the friendly sort. Each firm had a lot of respect for the other, but it was good fun to get ahead for a little while.

David

GMB 14th Jun 2018 9:30 am

Re: Oscilloscope bandwidth
 
Quote:

Originally Posted by Bazz4CQJ (Post 1051620)
Attached is a circuit from the 2008 ARRL Handbook intended to act as an adapter (convertor?) to enable a scope with a limit of 5MHz to work at 30MHz.

What a weird idea. The snag with doing this will be that anything other than a sine wave could look different to the original, so quite a misleading display. I think they had the narrow agenda of looking at carrier modulation, which will work OK.

GMB 14th Jun 2018 10:04 am

Re: Oscilloscope bandwidth
 
I just did the calculation of cut-off frequency for the Miniscope tube and I have to say that the suggestion that the frequency limit for the deflection plates is about the electron transition time seems wrong, or I am getting this wrong.

The electrons will be going at about 18,000,000 m/s so they can't be spending more than a couple of nS inside the plates. This seems to suggest that the limiting factor is much more about driving the plates than the electrons spending too much time there.

So I conclude that the limitation in this case is just that the Y plates are attached by long leads. Even so I would have thought that a decent Y amplifier might get it into the low MHz region so there should be scope for much improved bandwidth.

It seems to me that even the distributed amplifier systems are not about simple transition times but rather getting the plates charged up in the time.

turretslug 14th Jun 2018 10:37 am

Re: Oscilloscope bandwidth
 
I'm assuming that David (Radio Wrangler) was discussing the general outer limits of 'scopiness (the time-honoured analogue CRT variety, that is)- the limitations when dealing with hundreds of MHz, rather than hundreds of kHz. I know that 'scopes have always held something of a fascination for many, representing the optical interface between electronics and humans, and as such, they were eager embodiments of whatever was new and clever in electronic development. Our physics teacher had done something important and leading-edge at CERN before the ignominious come-down of dealing with oiks like us and it took little encouragement to divert a lesson into the esoterics of high-end 'scopes- he reckoned that delay-line tubes could stretch to a couple of GHz.

gingpeakin 14th Jun 2018 10:39 am

Re: Oscilloscope bandwidth
 
Was that W S Percival of Percival peaking fame? I think he was at EMI and patented the idea of a spiral anode that acted as a transmission line.

Ging


All times are GMT +1. The time now is 3:23 am.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.