Batboard

I have a UHF CDM1550/GM360 which is working OK for weak signal working after lowering the squelch threshold and tweaking the Ref Osc onto frequency ie. +/- 400Hz.

I am thinking about maybe tweaking the Front End filter softpots for even better sensitivity.

I'm interested in whether anybody has managed to improve on the factory Front End filter settings?

I have previously aligned other makes (Icom/Tait) and found that the UHF models were very very broad as far as Front End alignment was concerned - small changes in the softpot values bought about no real change in sensitivity.

The Icom radio actually had an automatic align function - each time you ran the auto-align function , you got different values for the Front End softpots - but no noticeable difference in sensitivity!!

So is it worth carefully tweaking the CDM1550 Front End softpots for max sensitivity?

Any thought?

Like the Icom you've tweaked, the CDM front end varactor filter tuning is very broad. I've never found a worthwhile increase whenever I've tuned the CDM or HT series, maybe 0.01 uV or 0.02 at most. That's why the tuner help advises you to actually tune at one frequency and then calculate the remaining softpot settings and then write them to the radio.
That being said, remember that 10 or 12 dB sinad sensitivity is as much a function of the noise figure of the first RF amp as well as the loss in the varactor-tuned front end filters. The VHF CDMs, for example, are speced at 0.35 uV @ 12 dB sinad, but all the ones I've seen are better than that by 0.07-0.1 uV.

Al - thanks for the interesting feedback.

I've often wondered why Moto use such pessimistic sensitivity figures (0.3/0.35uV for 12dB SINAD) when the actual sensitivity is always much better.

Interestingly, the CDM1550 was the 1st time that I've seen a "typical" sensitivity figure mentioned in the service manual (0.22uV) as well as the "official" figure (0.3uV).

Some of my most sensitive Moto radios are dedicated 12.5KHz radios - I understand that the narrower bandwidth has around a 3dBm benefit for sensitivity. This is noticeable with my MCS2000 - when I toggle between 12.5 and 25Khz spacing there is a very noticeable improvement in sensitivity on the 12.5Khz spacing (as well as higher RX audio gain/lower TX dev). However, there is very little (if any) sensitivity change with the CDM1550 when I toggle between 12.5/25Khz spacing - maybe the radio uses a "universal" compromise filter. I've seen this before with some Tait radios - a "universal" 10KHz IF filter is used for both space settings - which seems to work OK.

g8tzl2004 wrote:I've often wondered why Moto use such pessimistic sensitivity figures (0.3/0.35uV for 12dB SINAD) when the actual sensitivity is always much better.

Scenario #1:
"My radio is spec'ed at 0.35uV for 12dB SINAD, and it is measuring at 0.2uV".

Scenario #2:
"My radio is spec'ed at 0.20uV, and it is measuring at 0.22uV."

Question: Which scenario is likely to result in a call to the manufacturer?

Alternatively ,

"I bought a Vertex radio because it is more sensitive - quoted 0.25uV compared to the deaf Motorola which is only 0.35uV"

But the Motorola is actually 0.13uV!!!!

A lost sale?

Since the CDM and Waris HT series do, in fact, switch both sets of IF filters when going from 25 Khz to 12.5 Khz and in theory at least you should see about 3 dB better sinad sensitivity when you halve the IF bandwidth(testing at 60% of peak deviation), I've always wondered why Motorola specified the 12.5 Khz sinad at a higher uV value than the 25 Khz one. That seems to be the case on most if not all of M's mobiles and portables with switchable IF filters.

I was once told that narrower filters have a higher insertion loss - which might explain why some people think that a narrow filter set is deafer!!! BUT I've always found that narrow filter sets are usually more sensitive!! I guess sensitivity measures should also state what IF filter/spacing has been used for the measurement but you rarely ever see this.

I am currently programming a UHF Simoco handheld where there is "true" 12.5/25KHz filter switching - the difference in sensitivity is very noticeable - I'm sure its actually more than 3dBm!!

I never really notice any "filter rattling" when using narrow filters when people are TXing on wide dev - so the best thing for max sensitivity is to use the narrow filters and just increase the TX dev - maybe to a "compromise" 3.5KHz.

I've often seen mention of "12.5 KHz IF filters" but I think the actual IF filters for different spacing is as follows:

12.5KHz spacing - 7.5KHz IF filter
20.0KHz spacing - 12.5KHz IF filter
25.0KHz spacing - 15.0KHz IF filter

So the 12.5khz IF filter usually applies to 20KHz spacing. Narrow 12.5KHz spacing uses narrower 7.5KHz IF filters.

Whether the 12.5 Khz filter insertion loss is greater than the 25 Khz shouldn't make any difference....the filter loss attenuates the noise as much as the signal. What's really more important if you're doing engineering work on IF chain design is the individual filter's noise bandwidth rather than the manufacturer's specified (3 dB) bandwidth. Unless the filter is a brick wall filter, the noise bandwidth is always larger than the specified bandwidth. The noise bandwidth determines the ultimate (C+N)/N at the IF predetection point. And the (C+N)/N is one of the factors that determines the 10 or 12 dB sinad sensitivity. The other factor is the post-IF distortion introduced by the quad detector and audio amp chain. You may not hear a readily audible difference when the Rx carrier starts banging the sides of the IF filter(resulting in sideband attenuation), but it shows up as increased AF distortion and it's easily measurable. If a mfr. is going to spec.(and guarantee) such things as AF distortion it has to be measured on a typical prototype and then worst-case estimated deviations have to be added into the "typical" measurements to determine the number to write into the spec.

I've also often wondered how the Front End and Squelch softpots values are interpolated after you have aligned the radio at the 7 test freqs.

I've always assumed that the radio firmware uses a "line of best fit" (linear regression) technique based on the test freq values to work out the ACTUAL softpot value for any particular frequency.

So any small error in one particular Test freq value will be "averaged out".

Likewise, if you want to lower the squelch , I've always assumed that you need to do it for ALL test freqs. I've heard of people just lowering the squelch softpot closest to the intended frequency of operation. Assuming my thinking is correct , if you just do the tweaking on one softpot , this will require an enormous movement in order to move the overall "average" softpot line of best fit - maybe I'm wrong!!!

g8tzl2004 wrote:I've also often wondered how the Front End and Squelch softpots values are interpolated after you have aligned the radio at the 7 test freqs.

I've always assumed that the radio firmware uses a "line of best fit" (linear regression) technique based on the test freq values to work out the ACTUAL softpot value for any particular frequency.

While I don't know for certain, I would very much doubt that the radio is using anything that fancy - likely it is just using a piecewise-linear approximation (given a table of freq:pot pairs, locate the 2 table entries on either side of the frequency of interest, and linearly interpolate between the 2 points). Over the limited frequency range a normal radio supports, and given 7 points along that range, and given that the response curve for such things is usually pretty simple, the delta between a PWL and a linear regression is going to be negligible.