Compensating for VFO drift
 

I have built a quick VFO as a test case for measuring VFO drift. This is a simple Spectrum 2u6L coil, 27pF NP0 capacitor in parallel and 2x 330pF NP0 caps in a colpitts arrangement. There is a 2n3819 driving this and it uses a 78L06 as a voltage source. Driven from linear supply and monitored with a TF930 counter connected to the computer using some software I have written. This is all inside a cardboard box and the room temperature is a constant 19.2 degrees.

Getting slightly horrible drift on this as illustrated below:

Attachment 156179

X axis is time in seconds. This is approx 34 minutes.

Does anyone know of any canonical references for tackling this problem as I evolve this VFO into something usable?

Re: Compensating for VFO drift
 

The Oxley "Tempa-trimmer" being a thing of the past, you can achieve the same adjustable-temperature-coefficient effect by using a small differential-capacitor and one negative and one positive temperature-coefficient capacitor (both identical actual value).

Must admit, these days I generally avoid free-running VFOs unless they're coerced into stability by something like a 'Huff&Puff' sample-and-hold circuit.

Re: Compensating for VFO drift
 

Interesting idea. Will try that. This is certainly quite frustrating trying to fix. Perhaps not bothering is the best option :)

I've looked at huff-puff before. Definitely an option for this so might go down that route. I want this to be very stable as I've got a very narrow audio bandwidth filter on the other end of the receiver for CW reception so it's frustrating when it drifts out of the pass band. Grr!

I have looked at a proper DDS VFO. In fact I built a Si570 one from scratch but the thing drains 80mA of current continuously when receiving versus 1mA for this!

Re: Compensating for VFO drift
 

Radio books from the 1960s or early 70s may be the best source of information on building stable VFOs. Before that people were more likely to be 'rock-bound' (crystal osc). After that synthesisers started appearing.

My guess is that much of the drift comes from the coil. Reducing the oscillator amplitude might help. Good VFO coils were usually purpose-built.

Re: Compensating for VFO drift
 

Has per Dave’s suggestion, the ARRL handbooks from that time are a wealth of information.

Re: Compensating for VFO drift
 

I've got 1988 ARRL handbook, W1FB design notebook and Solid State Design for the Radio Amateur here and that's a little sparse on details. A good idea however. I will have a read through some of the material again.

I think you are right with the inductor. The other caps are all NP0/C0G so I wouldn't expect this level of drift. I will try replacing it with a material 6 toroid (T37-6) and a trimmer capacitor and measure stability again.

Worst case, I'm getting a lot of data on Spectrum coils :)

Re: Compensating for VFO drift
 

6Tanuki has nailed it.

Back in Pye Vanguard days I built an 11MHz VFO for use on the receiver. Of course, its output and its drift were multiplied by a factor of 12 at the final frequency. Also, the set would start at room temperature and head to somewhere well over 40°C, and then heat and cool further depending on whether it was on transmit or receive. I eventually got a tolerable outcome, but only after spending some time time (several days) playing at being "Tempco Man"! ;D.

My recollection is that there are still some neg temp coef caps around in the form of NOS tubular ceramics and polystyrenes too. Partridge electronics could be a place to try.

Torroids....really the way to go for a stable oscillator, why so ????

B

Re: Compensating for VFO drift
 

Turns out I actually did something ****** stupid. When I was dead bugging this, there was a solder spike across the series capacitor between the tank and the oscillator. So what's basically been happening is the coil has been happily sitting there warming up due to the DC bias across it. Tapped it with the iron and it's nowhere near as drifty as it was. Collecting data for this cock up now.

Well there goes another 3 hours of my life and another bald patch scratching at my head :)

Re: Compensating for VFO drift
 

Differential capacitors can also be hard to find.

If you look in any post-1995 ARRL handbook, in the oscillators chapter, there's my easy-to--adjust thermistor/varactor trick circuit

David

Re: Compensating for VFO drift
 

In 1960's and 1970's editions of the RSGB handbook, there is an excellent piece by G3PDM on a temperature compensated VFO for his receiver design, but it's based on the tempatrimmer.

David

Re: Compensating for VFO drift
 

I suppose instead of using a differential capacitor you could use a couple of trimmers instead. Would just make adjustment slightly more difficult.

Re: Compensating for VFO drift
 

You need to get the drift down, then compensate for the final tweak.

Toroids might not be the best option. Most good VFOs used an air core coil.

Re: Compensating for VFO drift
 

The differential capacitor is the easy way if you have one, but if you have not, you can still improve things a good deal even through trial and error. Many (most) components in a VFO have positive tempco, so a starting point is to try squeeze in any neg tempco you can. My memory may be failing me but re the cores of inductors; preferably air (?) (as our HRO's tell us), but next best is brass (?), preferably just inside the edge of the coil so just enabling some small degree of RF adjustment.

Regarding the room temperature; well it may look steady on a thermometer, but it ain't that good. As soon as you go near that 3819 an additional 400W(?) of heat arrives on the scene.

Enjoy the experience; crafting a good VFO is a rewarding challenge. Of course, access to a really good soldering iron (like the Aoyue 937) is a huge help ;D.

B

Re: Compensating for VFO drift
 

There is a lot of good advice and help in Chapter 10 of Crystal sets to Sideband by Frank W Harris, the rest of the book is very informative too.

Hope it may help.

T

Re: Compensating for VFO drift
 

Thanks for that suggestion - the whole book looks rather good!

Re: Compensating for VFO drift
 

Ok spent the evening on this, between being stressed about a pizza delivery, and some results finally. I have basically read all the books and articles suggested and come up with a Hartley oscillator instead from a W1FB design.

Dug around in the junk boxes and found a whole bag of 1971 dated 2n4416's and some enameled wire and went to town on winding my own coil. It is wound on paper on an X-Acto knife (exactly .3 inches in diameter) with masking tape over the ends and a fetching colour of nail polish stolen from my eldest to finish. tapped 1/4 of the way up. Measured at 4.7uH, a little higher than the 4.4uH predicted.

Is it stable? Considerably better. After 5 minutes of settling it drifts about 20Hz/minute which is pretty good. It is currently residing in a cardboard box for lack of something better to stick it in. Because it works pretty well, I'm going to reduce all the coupling capacitors considerably from 22pF down to about 10pF and see if it'll still start. This should give better load isolation.

Work in progress (horrible looking I know):

Re: Compensating for VFO drift
 

Final statistics for the night:

Version 1 - Colpitts stolen from G3RJV "sudden". Spectrum 2u6L coil + 2n3819 @ 7MHz

Version 2 - Colpitts fixed defective coupling capacitor. Specturm 2u6L coil + 2n3819 + buffer amp. @ 7MHz

version 3 - hand wound air cored coil + 2n4416. Design as per W1FB design notebook. @ 6MHz

Not sure why version 1 and 2 were as miserable as they were!

Attachment 156216

Next plan is to crank it down to around 3MHz and see how stable it is there. Then mix with a 4.915MHz crystal oscillator and bandpass it to give a stable 40m VFO. And then see if I can add varactor tuning to the lower frequency oscillator. And then huff puff no doubt.

Re: Compensating for VFO drift
 

Any chance of getting your hands on an old crystal oven casing from a commercial radio transmitter?
GE, RCA, and Motorola all used them, as well as some old General Radio gear.

The ovens were insulated & had a heater in them to hold everything inside at a certain temp.

Re: Compensating for VFO drift
 

Interesting idea. Ultimately one of the reasons this is being built though is that it is going in a portable QRP transceiver and the current DDS VFO is consuming a lot of current (85mA on receive). Objective is as low current as possible here. Ergo I’d like to avoid the current consumption and weight of an ovenised oscillator. For the purposes of development I have considered building a thermostat controlled environmental chamber however so that might fit the bill for that. I can plot drift vs temperature then.

Edit: I will defer thinking about this too much though for now as I’ll never finish the thing!

Re: Compensating for VFO drift
 

Seeing as you're planning to up-convert anyway, is it possible to invert the drift with your upconverter and then mix it with the non-inverted drift to get a stable version of the frequency you need?
Graham