Advance LF Generator Type SG66

[Radio Wrangler]

Tiny bead thermistors are still made for use in RF/Microwave thermistor mounts for standards lab grade RF power measurement but they're usually 100 Ohm at working temp and aren't in an evacuated bulb. So if you come across those, they don't work.

Real R53s are still available, just supplied in a nice case with the rest of the wien bridge included....

30 some years ago i was playing with some serious audio amplifier designs and I simply couldn't get good enough signals. In the end I built filters specifically for each of a group of frequencies. Doing low distortion filters wasn't trivial, either. One easier way is to combine two signals and do an intermodulation type of test for linearity. The finished amp could be run 1dB below clipping and I could show that all harmonic terms were more than 100dB below the fundamental. Doing the same thing with a pair of equal amplitude tones, with the peak envelope power 1dB below clipping, all intermod products were more than 100dB down. I could use tones anywhere in the audio range. I was being silly. Only magazine contributors have hearing in that league I was just playing, but I had to build the test gear to verify it with. Can I hear the difference? I don't think so. Did I have fun? Yes, lots.

OK, Dave, you have a box with some valve sockets, a ganged capacitor and a high-Z wien network. Hit the BAMA manual archive for the HP200C or D. They were the version that used a lightbulb "Bill's Bulb" and miniature valves. It'd be easier to nick Bill's circuit than to find another R53, I think, too many people have been looking for too long.

Cheers
David

[davidgem1406]

Hi Pete,

Thanks for the link I have saved the pdf of the page and will look into it.
I have the glass bead thermistor on its way so will try that before going any further.
I am also going to rewire again but change the layout that way I do away with the links between the 2 valve bases, V2a becomes V1b and V1b becomes V2a. I am hoping it may help the situation.

Hi David,

I will have a search for the HP manuals, I had thought of going that way and using a lamp instead, or even rebuilding the bridge using solid state but I don't know if I could drive the amplifier to obtain the correct voltage output. There is plenty of info for solid state Wien Bridges out there but not for valves, also they all seem to use a pot for tuning and that just wont go for this.

Thanks
Dave

[XTC]

Originally Posted by Radio Wrangler

Real R53s are still available, just supplied in a nice case with the rest of the wien bridge included....

If you are an instrument company in China with an initial firm order of 20,000 and an RA53 is the best way to go, it won't be much of a problem getting one made.

If you are Agilent etc, things have moved on from the RA53 and had at least 20 years ago.

I restored a Rapid basic AF sig gen and it had an RA53 type thermistor in a metal can looking a bit like an OC81. I'm pretty sure Rapid sold it until the past few years.

RA53s were also part of the OU scope and function generator which they provided with their electronics degree courses. I've only seen one and it was in a sad state about to go to the dump, but I rescued the RA53.

Originally Posted by Radio Wrangler

30 some years ago i was playing with some serious audio amplifier designs and I simply couldn't get good enough signals. In the end I built filters specifically for each of a group of frequencies. Doing low distortion filters wasn't trivial, either. One easier way is to combine two signals and do an intermodulation type of test for linearity. The finished amp could be run 1dB below clipping and I could show that all harmonic terms were more than 100dB below the fundamental. Doing the same thing with a pair of equal amplitude tones, with the peak envelope power 1dB below clipping, all intermod products were more than 100dB down. I could use tones anywhere in the audio range. I was being silly. Only magazine contributors have hearing in that league I was just playing, but I had to build the test gear to verify it with. Can I hear the difference? I don't think so. Did I have fun? Yes, lots.

Been there with a JLH design which produced 0.002% THD as specified and as measured with an 8903 with the 80KHz filter in. Only 0.006% with the filter out. A Marconi TF2104 consistently showed 0.002%. No thermistor in either. CdS/LED combinations. I abandoned the project as a dangerous obsession.

As I recall, tests in the 50s showed that even order distortion wasn't detectable by anyone below 1%. Then there's the distortion introduced by speakers.

Originally Posted by Radio Wrangler

OK, Dave, you have a box with some valve sockets, a ganged capacitor and a high-Z wien network. Hit the BAMA manual archive for the HP200C or D. They were the version that used a lightbulb "Bill's Bulb" and miniature valves. It'd be easier to nick Bill's circuit than to find another R53, I think, too many people have been looking for too long.

You have to ask yourself what you are doing; restoring, having a bit of fun with junk which cost a few quid and learning something, building the ultimate low distortion oscillator. If it's building the ultimate low distortion oscillator, valves are a bad starting point and it's probably easier to compromise your principles and just buy a pretty good one which will be well capable of audio testing.

I agree that using a lamp is the way to go, but I'd be more inclined to mod the existing circuit than copy Bill's design.

I think the thermistor was a way of evading Bill Hewlett's patent.

Pete.

[Radio Wrangler]

Hi Pete,

Bill's patent was in 1939, so it would have expired before the 1960s. I think the bead thermistor was designed to give a bit more appropriate time constant than the bulbs did. Still, the bulbs worked fine, but the 200 series oscillators weren't ultra-low distortion jobs.

We must have had similar thoughts about it being a dangerous obsession. I did my amp to drive a pair of transmission line speakers and it's sat in the living room for the past 30+years, just working whenever I want it. There's been no need to change anything, though I did wear out the drawer of a posh Sony CD machine over 25 years. Oh, a small run of the amps was made and there's a dozen or so of them kicking around. HP made PCBs as a training exercise.

On the ethos of restoration. The Advance generator could be stuffed with a different circuit with a bulb not an R53. That makes it a working unit for fixing other things.

If the Advance generator is seen as a restoration project itself, then getting an R53 means ending the possibility of restoring *its* donor. I suppose I'd happily see a common type sacrificed to repair, say, a Radford low distortion oscillator.

Bills design runs highish power. It has to to drive the bulb. Modding the existing circuit would entail adding power stage of a few watts. The original octal version used a 6V6 class device.


Cheers
David

[davidgem1406]

Hello all,

I think this is now getting to involved for what I want to do, all I intend to do is to get this Advance working as it was meant to be. I wont be needing to test HiFi amplifiers, maybe the odd audio amp now and then.

The thermistor is a problem but as I read it that is only used to stabilize the amplitude of the waveform by controlling the gain via feedback.
As for replacing it with a lamp I don't have a clue where to start the modifications for that one.

I am more concerned at the moment with the problem that the frequency is basically running out of control since I rewired the oscillator to the circuit in the manual, and that should not be happening as all the wiring is correct and I replaced all the resistors at the same time. The existing frequency range selection components are correct to the manual as is the rest of the circuitry.

I will redo the layout as I mentioned earlier and see what happens then, if it still has a major problem then I will probably return it to as it was previously, at least it was running close to what it should be.
It seems somewhat strange that the circuit from the manual should be producing this sort of problem.

Thanks for all the input so far.

Dave

Well, well as I was writing this the glass bead thermistor just dropped through the letter box. (Post Number #39)

[XTC]

David,

I've seen and made op amp sig gens with a bulb. it doesn't have to glow. Admittedly, the one I built used an LM386. I forget what the performance was exactly, but it was quite respectable, especially for a simple, cheap circuit.

I've no doubt the thermistor needs less power than any lamp. The lamp is cheaper. The last time I saw RA53s for sale in a catalogue, they were £6 and that was a long time ago.

As for donor equipment, well, you come across all sorts, some simply not in a state to restore at all. it's a question of recognising it. The OU scope had no cover, missing parts, had been left in the damp for years and had a shot tube. The only thing worth recovering was the RA53.

Barny Oliver designed a hi-fi amplifier which was not quite an official HP product.

http://www.hpmemory.org/news/barney_...ey_page_00.htm


Pete.

[davidgem1406]

Hello Pete,

Yes there are plenty of solid state circuits on the net for wein bridge oscillators but they all , so far anyway, seem to use a pot for tuning and that precludes using any of those as there are only 190 degrees of rotation available on the tuning scale, also I'm not clever enough to redesign one to use capacitor tuning.

I will keep looking.

Thanks for that link it looks like an interesting article and I will get to read it in stages as things will allow.

Dave

[Radio Wrangler]

Originally Posted by XTC

Barny Oliver designed a hi-fi amplifier which was not quite an official HP product.[

Yup, know it well. Very smart cookie Barnie. Involved with all sorts of things like SETI (serious DX!) and founded HP labs.

Mine was several years later. I've got got a couple of boards here. Double sided, copper-nickel-gold process, HP logo and a training number Datecode 2110. So week 10 1981.

Hasn't been touched since.
Hasn't needed to be

Oooops, find some wood!

David

[davidgem1406]

The next chapter in the Advance SG66 story.

As I said I would I have now rewired the oscillator panel using a different layout.
I used the glass bead NTC thermistor, this is a 100K device, that I purchased. I used it because when I measured the original thermistor it measured 100K, unless I made an error. However I looked into the RA53 and found a data sheet covering the RA range of thermistors and the RA53 is quoted as 5K.
That now makes me wonder what the resistance of the R53 actually should read, all at ambient temperature of course.

Anyhow the generator is working far better than the previous rewire.
Ranges 1 to 4 produce a reasonable sine wave and are close to what they should be in frequency. The amplitude across the 4 ranges is within 1.5V ptp, not to good but way better than last time.

Range 5 the highest is another matter the sine wave is awful, I attached a picture, easier than trying to describe it. I have to say that there is a cathode decoupling capacitor (C9 100pf) where I used a suflex capacitor as that is all I had other than mica. The suflex was easier to fit in a tight space.
Maybe that is not good for the range?? although it is only 50KHz to 150KHz.
The set pot for the low end of the range is changing the amplitude and not the frequency.

I'm not sure that the thermistor is the right one to use, but it is a step in the right direction.

Dave

[XTC]

The point about the RA thermistors is that they are a tiny bead of NTC material on two fine wires in a vacuum. The bead is probably operating at a temperature between 50C and 100C making it independent of room temperature and the vacuum helps control the time constant and maybe protects the bead from chemical attack. The rate at which the bead heats and cools is an important part of the regulating action, as is the non linear voltage/current curve.

www.thermometrics.com/assets/images/ra.pdf

A quick look on the thermionics website shows nothing similar in their current line.

In many ways a bulb filament is the same, but is PTC and takes a bit more power.

A disc thermistor may be self heating, say to deal with current inrush, but will be a totally different beast I suspect the thermistors you bought are intended for temperature sensing and will give little advantage over a resistor. In my view your scope trace shows the gain set to well over the required value to just sustain oscillation and you are seeing clipping.

The Suflex capacitor due to its wound construction, may have a troublesom resonance. I'd try a disc ceramic.

Pete.

[davidgem1406]

Hi Pete

Yes I already have that data sheet it is the one I was referring to.
If the R53 is also a 5K device then the thermistor I have in circuit now is way to high in value at 100K.

Replacing the thermistor with a lamp is an option if I knew where to fit it, the mods involved and the voltage / current of the lamp.

The waveform shown is for band 5 only the other 4 produce a sine wave that is reasonable. Adjustment does nothing to improve band 5, just degrades the other 4.

The suflex will be replaced as soon as I have some ceramics of the correct value. Also I will get a 5K NTC glass bead thermistor to try.

Thanks Dave

[davidgem1406]

I'm thinking that maybe I have solved the thermistor problem.

I have found a source that has STC B54 available and it is designed for the purpose has the same construction as the type RA53.
Not sure how they wire in as they are a 4 lead device and no info on the data sheet as to how to connect.

The data sheet is attached.

I spent all day Wednesday hunting and it is now 1am Thursday.

Dave

[Radio Wrangler]

This is a thermistor with a heater included. They get used in several applications the simplest is as a variable resistor... Drive the heater to make the thermistor vary, make sure the thermistor self-heating is negligible, the other main use is in true RMS AC (orRF) to DC converters, usually used in matched pairs where the resistance of two thermistors is compared and used to steer the power fed to one heater to therefore match the power fed by an unknown signal to the other.

One of the models will very likely have the same bead as the R53, but there will be the added thermal mass of the heater...

David

[davidgem1406]

Hi David,

Thanks for that information, I now see how it would work.

If however the heater is not used then it would work in the same way as the RA53? wouldn't it!

On the other hand if the heater has to be driven to vary the temperature of the bead and the bead does the stabilizing as normal that then presents a hole new ball game, where to get the heater drive from? maybe of the oscillator output some how.

There are no options as to model the B54 is all that is available.

Dave

[XTC]

I'm sure this is the sort of thing that people have found in their junk boxes and raised threads on here about as mystery objects. I seem to recall it was said that they used to be used in telephone exchanges by the hundred for amplitude control. I don't know whether they were similar or a totally different pattern.

I'd say it was worth a go if it's not too dear. You might have to modify the circuit a bit.

If you drive it through the heater, I would have thought the time constant would be all wrong and you'd have a self-heating effect on top. I've never seen one and I'm not sure how the heater was linked to the bead.

Pete.

[davidgem1406]

Hi Pete,

I have 2 on their way to me, at £5.20 each, not bad. 2 just in case i kill another one, hope not though.

I will use it as though it was an R53 and see what the results are.

Dave

[Refugee]

I had a look at the PDF and it looks like an NTC.
I would look at the FET ones in the transistor circuits.
You can get FETs that work at high voltage these days and the circuit will not have a care in the world about the pot/capacitor tuning issue.
It only has to measure the voltage and lower its resistance a bit as the voltage rises.
The LED one might not be happy with the voltage.

[davidgem1406]

Hello Refugee,

Yes it is an NTC device same as the R53, that is what it should be.

The pot tuning issue was mentioned as I was looking at using a solid state oscillator as a replacement. All the circuits I have seen use a pot for the tuning and that wont go here.

The FET option could no doubt be used but I have no idea as to what mods would be needed to the existing circuit to accommodate it.
The thermistor is still the better option over the FET if the thermistor I have coming will work OK.

I have not seen anything about using an LED option?

Dave.

[Refugee]

It is a long time since i worked on something with one in it.
I was thinking that was what was making it hard to get.
A test could be done by insulating an ordinary one in expanded polystyrine at the expense of settling time.
Modern chips are likely to have the FET type connected to a pin where the timing capacitor is added.

[davidgem1406]

Hi Refugee,

Im not sure we are on the same wave length here. This is a signal generator using valves. The oscillator is a wien bridge that used the thermistor to stabilize the sine wave amplitude.
I destroyed the original thermistor, whilst working on the oscillator circuit, and they are all but impossible to get now. I just happened to come across the B54 whilst searching for other options.

The solid state circuit was just an option that I had looked at but the mods to the existing unit would be extensive, also it would need a circuit that used capacitors for the tuning not pots.
I think practicality says it is not really an option.

I did think that you meant the FET option as a replacement for the thermistor in the existing circuit, unless I am missing something here.

Regards
Dave