Modified Advance E2 Sig. Gen.

[David Simpson]

Calling all Advance experts. I've just acquired an E2 which looks as though it was built in the 70's, as inside its immaculate. Also, it has some semi-conductor & valve improvements. V1 is an ECC91(not an ECC82), V2 is ECC82(not a larger 6SN7GT), and V3 is an EZ80(not a larger 6X5G).
Off V3's cathode a wee s-cond. F/W rectifier circuit has been mounted on a little tag strip which supplies -200V to an additional veroboard circuitry containing a TO5 BFT44. That in turn takes a feed off V1's grid, then puts a fixed 7mV P-P RF o/p to a jack plug on the faceplate.
Did Advance make these improvements to this range prior to ceasing production of this 1950's designed sig gen? Or perhaps mod kits were supplied to owners ?
Any advice most appreciated, please.

Regards, David

[AC/HL]

It was a long production run. The valves probably went from current to maintenance types at some point, necessitating redesign. Is that socket for a counter, they would have been needed to keep pace with technology and keep it competitive.

[turretslug]

Ahhh, the VW Beetle of sig. gens..... Produced for ever and a day, works simply, unpretentiously and without fuss for ever and a day. I wonder if the designer owned/hankered after an AR88 when it came to the frequency display.

I'm sceptical that a mainstream manufacturer would have used that amount of Veroboard for such a simple and small-area circuit- or even any Veroboard, a postage-stamp of plain matrix would have been quite adequate and would have minimised stray capacitance. Speaking as someone who's tried sampling oscillators with minimum disturbance over the years and never quite got the point of Veroboard- certainly for an application like this.

Deriving a high -ve voltage and using an HV transistor (with the implication of low current gain and seemingly unnecessary circuit loading) for such a low-amplitude output are also curious aspects- though I'm quite prepared to accept that there is genius or green-fingers at work here that has flown over my befuddled head.

Assume this is for a counter or wave-meter output, but a jack socket is odd for RF!

All seems like a diligent hobbyist who did things his/her own way.

There's an E2 here that sounds like an interim job, with ECC91 (paralleled) oscillator but 6SN7 and 6X5GT- I suspect that there would have been a lot of "ongoing revision" with a long-running simple stalwart like this, not to mention the changing surplus market parts availabilty! They are very well sealed when screwed up and getting them open can be injurious to finger-nails, I wouldn't be surprised if most have a "could eat your dinner off it" aspect inside, unless grievously neglected

[Station X]

Despite having a couple of synthesised SG's with digital displays and coverage into the Gigahertz region, I still have an Advance P1 SG to fall back on. Ideal for domestic sets. They're also easy to fix if they go wrong.

[turretslug]

Is the base of the BFT44 DC-connected to the ECC91 grid?- that would allow the running -ve grid voltage to bias it, a quick search shows it as relatively low input capacitance and high hFE. So, an appropriate input resistor would allow DC bias and AC stand-off, with considerable signal attenuation- but there would be several volts amplitude on the oscillator grid and the output might only need to be a few mV, as found.

So, maybe crafty after all.

[David Simpson]

Many thanks Guys for your interest. As well as the 7mV p-p from the jack socket, there is 1.3V from the "Full RF" Belling-Lee skt., and 320mV max from the attenuated o/p. Strangely, apart from the normal RF o/p's having the 400Hz 30% modulation envelope when on Int. Mod., the jack skt's o/p is 400Hz with the RF superimposed on it !
Seemingly, this sig. gen. spent all its working life(but hardly used) from new in a University lab, but was rescued from going into a skip some 20 - 30 years ago when it was declared obsolete. Then it was never used & kept in a nice old dry cupboard, till a couple of weeks ago.
Why the jack socket was used instead of a B-L 70ohm skt., or a BNC, for RF is a puzzle. This 7mV p-p aux. o/p is hellish low for a freq. counter i/p. My Maplin MF100 needs a min of approx. 30mV from the att'n. O/P, the Dong Kwang I posted about a couple of months back, only likes the full RF o/p, and my wee Watson needs 20mV from the att'n o/p.
The add'l circuitry was definitely not done by a hobbyist, must have been a uni lab tech, if not done in the factory. Too high a standard of workmanship.
I'll gently remove the veroboard & try & work out the circuitry. So far, all I can see is a 110Kohm resistor in the feed from the grid. The resistors on the v/b are too minute to read the colour codes just now.

Regards, David

[David Simpson]

Hopefully the diagram of the veroboard's circuit will throw some light on the matter. The diode between the jackplug socket & the transistor's base is the tiniest imaginable, and the lettering barely readable.

Regards, David

[ms660]

Duff transistor?

Lawrence.

[Skywave]

Looking at that cct., it seems to me that the jack socket is intended as an input, not an output. Similarly, the connection to V1 grid is an output to V1, not an input to that transistor (BFT44). The cct. of the E2 caters for a modulated H.T to V1, the oscillator. That mod. freq. is fixed at 400 Hz (T2, C12 and half of V2). So I will hazard a guess that the purpose of that transistor stage is to provide for an external modulation signal to be fed in, whose freq. can be determined by the user, as opposed to the internal fixed freq. of 400 Hz. That ext. mod. might have been a square / pulse wave.

So whose handiwork was that then? Quite frankly, I don't know. But looking at the components, I suspect someone at the University; someone who was well-versed in the design and construction of analogue transistor circuitry.

Al.

[David Simpson]

Right enough, Al, I did wonder. I'll maybe try some o/p's from my Farnell FG2 into the jack skt. I wonder what the magnitude of the signal was. Quite low & not RF I suspect. Therefore not de-stabilising the osc circuitry. Otherwise, a jolly good Sig gen in excellent nick. Having a Marconi TF995A/5(& a 995A/2 spare), & limited collection space, I'm not really needing it.

Regards, David

[David Simpson]

Took a 2nd opinion from a retired electronics engineer chum - definitely a pulse modulation circuit. Going by the function gen i/p I used - sine, square both gave just about the same modulation result. As shown in picture - a 330mV p-p RF o/p reduced to a 160mV pulsed RF o/p. Must have been for a specific degree or post grad student project back in the 70's, then never used again. I'll just leave it in place, as it doesn't affect the sig gen's normal functions at all.

Regards, David

[Skywave]

"Pulse mod. cct." Yes, I thought as much when I saw the diode in the input cct. Glad to know that your question has been answered.

Al.

[David Simpson]

Moderators - Query resolved. Guess this thread can be closed.

Regards, David