Amstrad CB 901 do they drift in frequency?

[HeliJimbo]

Hi all,

I've just serviced a couple of Amstrad CB 901s I got for extremely cheaply.

It was a great result as both were Japenese and both had no mods in at all.

All that was cranked up was the deviation on both but apart from that, they were great, aside from the odd bulged capacitor etc...

Both of these units appear to drift around in frequency.

One minute they're perfect then a little later they're 150 Hz off.

They were both set when the unit had warmed up.

Is this common for these chassis?

Thanks.

Just to add to that, I also noticed that when switching to low power the frequency shifts by about 200Hz on both sets.

[Techman]

150-200 Hz is nothing when you're talking about millions of Hz!

Well within tolerance I would say.

Voltage drop around the circuit with power output difference is bound to cause slight changes in frequency.

All rigs drift!

[Andrew2]

I agree. Given that the channel spacing is 10 kHz and the RX IF bandwidth will be about 5 kHz, a drift of 200 Hz is neither here nor there, and is certainly 'on channel'.

[HeliJimbo]

Thanks chaps.

I know the 901 manual says it's +/-300Hz but I can't seem to find the source document for the tolerance.

Can anyone point me to this please?

I have tried looking.

[G0HZU_JMR]

In my student days when the UK CB craze was at its peak I used to repair CBs to earn some extra dosh on top of my student grant. I had a fairly decent frequency counter back then and was probably able to measure 27MHz to within about 20Hz. I still have the counter and it still performs well.

I do recall that many of the UK FM CBs were quite stable on transmit over a short term. By short term I mean cold start and maybe 10-20 minutes' operation. It was a long time ago for me to remember accurately but I would consider a 200Hz shift to be unusually high for the Amstrad 901 but not that significant in terms of the impact on performance as a CB. Note that because of the way these UKFM CBs are designed it isn't possible to achieve similar accuracy for the transmit frequency on every channel but the synthesiser error on transmit across 40channels is something like +/-10Hz if set for best accuracy on channel 19.

I'd recommend you check out how stable the main 9V regulator is over time. Does this drift up and down much as the CB drifts in frequency? These radios are 40 years old now so there's bound to be some degradation in the components so several things could be contributing to the drift. All I can offer is my recollection that these (Cybernet manufactured?) Amstrad 900/901 CBs were fairly stable in terms of the transmit frequency.

[HeliJimbo]

Thanks Jeremy, all taken on board and food for thought.

[G0HZU_JMR]

Thanks, I hope my input is useful but this was all a long time ago. How are you testing your radio? Are you comparing a cold start to the drift after heavy bursts of transmit for a long time where the radio chassis gets quite warm? I do remember these were the radios with the big green resistor inside. This resistor gets very hot on transmit. Enough to burn your fingers if you touch it.

I get the impression you are seeing a more random type of drift over a fairly short time and this would be quite unusual.

A 200Hz shift across low to high power also seems a lot to me. Hardly anyone used the power dropper on these radios so I'm not sure I would have ever tested this in terms of frequency drift. I think the power dropper on the Amstrad 901 was a little blue recessed push button at the back of the radio.

As others have already said this amount of drift doesn't really matter much for an FM radio but you did ask if the drift was typical for this radio type and I don't recall that it is. However, things might be different now as these radios are 40 years old.

I had a look through some old CB magazine scans and an early Amstrad 901 was tested by CB radio magazine in 1982 and they did a fairly harsh test where it was tested on a cold day vs a normal day (20 degF difference) and they saw 40Hz drift on transmit. I'm not sure how accurate their frequency counter was across those conditions as it would need a decent OCXO option fitted. I guess we have to assume they used an accurate counter at both temperatures.

[G0HZU_JMR]

I still have a lot of old UK and export CB radios and lots of CB spares up in the loft and had a rummage up there this morning. The nearest I've got to an Amstrad 901 is a Fidelity 2001FM. This has the same circuit board inside as the 901. I've also got a fairly modern Anritsu frequency counter here that has a decent 10MHz OCXO reference and this is usually within 0.03Hz of 10MHz after warmup. So I can measure the frequency quite reliably.

I just powered up the Fidelity 2001FM on channel 19 and the Tx frequency was 27.781238MHz on full power. This was a -12Hz error on a cold start (although it is quite warm today). With the -10dB power dropper button pressed the frequency changed by just 7Hz to 27.781245MHz.

Back up on full power the Tx frequency went up by just 2Hz when changing the PSU from 13.8V to 12.0V DC. This seems very stable to me. I'll leave it switched on and measure it at 13.8V DC at several times during the day. It has already drifted 2Hz in the time it took me to do the above tests as my workroom is quite warm today and the CB itself was cooler than the room when first brought in and powered up and tested. There's also going to be some warmup drift once I leave the radio powered for a while.

[G0HZU_JMR]

After half an hour warmup on receive only the transmit frequency was 27.781254MHz on channel 19. This is only a +4Hz error compared to where it should be. I then tried transmitting a few times and it drifted up about 1Hz.

I then tried transmitting continuously for 7.5 minutes and logged the counter with a PC. It drifted up about 6Hz in this time to about 27.781260MHz and the radio chassis was just becoming warm to the touch. I left it several minutes in receive and measured again and it was still at 27.781260MHz.

I then transmitted again for a steady 7.5 minutes and the radio chassis was then very warm at the back but the drift was less this time and it reached 27.781263MHz. After a few minutes' stabilisation time in receive mode it drifted down 1Hz to 27.781262MHz.

So if I measure the drift after the initial 30 min warmup in receive mode where it was at 27.781254MHz, it drifted less than 10Hz even with 15 minutes' of continuous transmission time.

Worst case from a cold start it drifted 25Hz from 27.781238MHz to 27.781263MHz once the chassis got quite hot after the 15 minutes spent in transmit mode.

I tried the power dropper again and the drift was +6Hz when in low power mode. This is a lot less than the 200Hz you are seeing. I hope this is useful info for you although this Fidelity 2001FM has had an easy life compared to your radios. It still looks almost new with no marks on it and the mic is still shiny with no scuffs on it. It is completely original and as new on the inside.

[Techman]

Those 9 volt(ish) regulators in these rigs often don't cope very well with large power supply voltage swings. It's never noticeable on FM or AM rigs, but it's generally what causes the characteristic 'warble' that's heard on SSB on sets that have this mode of operation. The frequency drift between power settings is being caused by the main power supply not being up to the job, or bad volts drop in the wiring, fuse holder, circuit board print etc. and the fact that the 9 volt regulator within the set is not coping with this line voltage variation. So it's either a fault with the main power supply or its connections, or it's the regulator and its associated circuitry that's struggling, possibly due to aging components. So from what's been said, it looks like the rig is capable of better, but whether it really matters is another story.

[electronicskip]

I must try my 901 and see if i get any major drift ,ive got a decent frequency counter so will post the results when ive powered it up.
On the plus side , my 901 has always behaved perfectly and for a cheap CB i could never fault the performance.
(must stop buying all these CBs) but hopefully will pick up one or two at Retrotech.

[G0HZU_JMR]

Hi electroniskip, it would be interesting to see some more results if you can test out your 901.

I tried my Fidelity 2001FM several times during the day and the transmit frequency gradually crept up to 27.781267MHz but this was mainly caused by the increasing temperature in my workroom. It is quite hot in here at the moment.

As techman says these radios rely on a basic 9V regulator circuit and sometimes the pass transistor or the zener diode gets replaced with non-equivalent parts when they fail.

I also found a little Cobra 21X FM CB. This is a bit scruffy on the outside but looks original on the inside. This is one of the cheapo Maxon chassis radios rather than Cybernet. I just tried it on Tx on CH19 and it was about -200Hz off frequency. When I adjusted it to where it should be the trimmer seemed very sensitive so I'm not sure how stable it will be over time. However when I tried the power dropper switch the frequency changed by less than 1Hz! I think this is a genuine 1981/2 built radio just like the Fidelity 2001FM. I think the Amstrad 901 had a longer production run and the quality wasn't as good on later radios. I remember there were some brown coloured clones of the 901 that were made by a different manufacturer.

[HeliJimbo]

Hi chaps,

I am following and have taken away that I think there is some kind of minor issue.

Lots of testing and tweaking today.

I have now tweaked the 10.240 on both units and from low to high power there is not much difference so that's good.

The voltage on the collector of the 10.240 oscillator doesn't change much either when the frequency changes.

I would not class this now as a drift, but more of a 150 Hz or so change at different times.

Once calibrated, when I go back later it's off, and then sometimes it's back.

All electrolytics changed and dry solders checked for.

What I've learned from this post is that what I'm seeing isn't normal so I'm cracking on with more fault-finding.

[Cruisin Marine]

I had a farmer as a client, had more acres than you could count down in Kent. Had about 11 or 12 901's that were used every day.
I used to check them out every 12 months or so, they were used in Tractors mainly, they hardly needed any re-calibration at all.

Damn great piece of kit, simple but reliable as hell.

[G0HZU_JMR]

The little Cobra 21X has been on for quite a while now in receive mode only. It showed 27.781263MHz on the first transmit attempt. So this was about a +15Hz drift because I didn't set the trimmer any more accurately than 27.781248MHz at the start of the test. After the long warmup where it drifted to 27.781263MHz, I then tried transmitting a 50% duty cycle for several minutes and the additional drift was less than 1Hz. I also tried varying the PSU from 12V to 13.8V and the drift was less than 1Hz.

The back plate of the radio got very hot after only 5 minutes of this so I didn't dare transmit for the full 7.5 minutes. Instead I transmitted for 30 seconds every 30 seconds. The drift was probably just over 1Hz during all of this and this was a bit of a surprise.

This drift performance is slightly better overall than the Fidelity 2001FM. The very basic receiver circuit in the Cobra 21X was very prone to overload back in the day to the point it was virtually useless in a busy town or city in 1982. I remember that it was popular to swap the ceramic filter for a basic crystal filter at the 10.7MHz IF on radios like this.

Back in the day it was the multimode export radios that drifted a lot as they typically used three crystals in the synthesiser and often one of them was part of a tuneable VCXO. Some actually used slide switches to switch in and out the band crystals and this was a bit of a bodge. Some of them were also very inefficient on transmit and this didn't help with the drift as the chassis could get very hot if used on transmit a lot.

[SiriusHardware]

This synth circuit was common to a lot of 27/81 radios not at all limited to this chassis so I am wondering if the crystals or the trimmer capacitors Amstrad's builder used were just not of the same quality / temperature stability as the corresponding parts used in Cybernet or Maxon chassis. Uniden favoured a different PLL IC). I suppose the way to rule it out is to 'borrow' the crystal and trimmer capacitor from the obviously very stable Cobra 21X either separately or together and see if they improve the behaviour.

There are commonly also one or more fixed ceramic capacitors between either end of the 10.240 crystal and 0V, and these would also cause frequency drift if they were more temperature sensitive than they should be.

[G0HZU_JMR]

I suspect that the reason the little Cobra 21X FM only shifts about 1Hz in low power mode or when the PSU is changed is because the crystal oscillator is integral to the PLL chip and the PLL is powered from a linear regulator IC that will have internal feedback to keep the 8V supply stable even when the transmit power is changed or the PSU voltage is changed. There is an RC cleanup circuit between the regulator and the PLL chip that might spoil things a bit but I think the 8V linear regulator must help a lot here.

[SiriusHardware]

Ah, I don't have my usual resources to hand so did not realise this chassis used a discrete 10.24 oscillator.