Heathkit HW-8 discovery

[m0cemdave]

These radios were sold from 1976-1983 so I presume are sufficiently vintage for this message board.

I have just been working on an HW-8 that was acquired some time ago at a rally. It had a couple of messy modifications that needed fixing properly, but it now works as it should although I still have to tidy up some of the original wiring.

I should explain for the uninitiated that this is a very simple but quite effective CW-only QRP transceiver for 80/40/20/15m, VFO controlled with about 2 watts output and using a direct-conversion receiver. Like many DC receivers this has a 2-stage audio bandpass filter to provide selectivity, switchable to "wide" with one filter and "narrow" with both in circuit.

One thing that has always annoyed me with HW8's - this is the second one I have owned - is the audio filtering. When switching from "wide" to "narrow", it is obvious that the filter centre frequency changes, and the gain increases too. This is really not what's wanted, especially when working weak signals with simple equipment.

Whilst looking at the circuit diagram, and just out of casual interest, I decided to look at the filter design. Each stage uses a single opamp in a bandpass filter, a standard circuit which appears in various textbooks. For example, page 82 of Hayward and DeMaw gives the formulae for the component values. For given capacitor values the filter response is determined by three resistors.

The HW8 manual quotes a centre frequency of 750 Hz and using that in the calculations, the component values in the first stage imply a voltage gain of 5 and a Q of 2 which is how the filter measures on test (actually 740Hz, close enough).

However those in the second filter don't make sense. R23 and R25 are correct for a gain of 1 (so the level should remain the same when switching to "narrow") and give a calculated Q of 2.34 for a 750Hz centre frequency, but the value of 82k for R24 makes no sense at all. Calculations indicate it should be 49k8.

Substituting with 47k provides the second filter with a measured gain of 1.02, a frequency of 740Hz, and a Q of 2.28. Now when listening to CW, switching the second stage into circuit has the effect of cleaning up the signal by reducing the bandwidth, with no shift in either the level nor the centre frequency. This is of course what one would have expected in the first place.

The HW8, which has a large following as a "Classic" QRP radio, is the subject of numerous internet articles and a popular booklet, with many modifications having been published over the last 35 years. I find it hard to believe that I'm the first person to have found this anomaly in the design but I certainly seem to be the first to have reported it. I wonder if somebody at Heathkit in 1978 made a mistake when specifying the component value?

[m0cemdave]

Sorry - a minor correction.

The final value of the resistor I substituted was 51k, not 47k. In my radio, it gave a better match to the first filter frequency although there will be some variation due to the capacitor tolerances.

R24 is rather hard to get at and I had to use a tweezers when changing it. If you want to try this modification you may find it easier to just add a 130k resistor across R24, on the back of the circuit board.

[Ian - G4JQT]

It may be worth posting your findings on the Heathkit Yahoo group

There may even be a dedicated HW-8 group - there certainly is one for the HW-9.

Ian

[m0cemdave]

I don't do Yahoo.

However I've just sent an edited version to Sprat, so maybe it will appear in there before long.

[Ian - G4JQT]

Another worthwhile mod for this rig is putting a 22uF capacitor across ZD1 which reduced the receiver noise considerably.

Ian