MTR 2 gain mod

My MTR2 now with a touch paddle

My MTR2 with a touch paddle

Part 1 (Dec 2014) – Simple gain mod

I’d read on the AT-Sprint Yahoo group about a simple gain mod for the MTR 2 radio that a number of people recommended to improve the sensitivity of the receiver. I decided to give it a try and also try and measure the difference rather than relying on subjective evaluation.

To start with, I measured the Minimum Discernable Signal (MDS) and the 20dB S+N/N levels to provide a baseline. Here’s the results across the 40/30/20m bands (which is what my MTR 2 is built to cover).

Frequency [kHz] Input level [dBm] Audio [Vrms] Detectable above noise?
7030 Noise only 0.0038
7030
-117 (S2) 0.0042 Yes – Minimum Discernible
7030
-88 (S6.5) 0.0399 Yes, strong – 20dB S+N/N
10116 Noise only 0.0039
10116 -117 (S2) 0.0049
Yes – Minimum Discernible
10116 -92 (S6) 0.0399
Yes, strong – 20dB S+N/N
14061 Noise only 0.0038
14061 -123 (S1) 0.0039
Yes – Minimum Discernible
14061 -89 (S6.5) 0.0400
Yes, strong – 20dB S+N/N

The test set up was as follows:

  • Signal generator was a N2PK VNA with a nominal output level of +4dBm. This was attenuated by 20dB and then measured as a reference level with a calibrated mW power meter.
  • This signal was then fed to a 0-1 GHz HP 355D 0 – 120dB 12 step attenuator and a second 0 -1 GHz HP 355C 0 – 12 dB 12 step attenuator
    • Whilst these are not calibrated, they check out well against my calibrated mW power meter
    • The series combination gives me 0 – 132dB attenuation in 1dB steps
  • Starting at a reference level of -17dBm, I could generate a test signal down to -149dm
  • Whilst the calibration of this setup is certainly not NIST traceable, it should be accurate to within 1-2dB

MDS is a rather subjective measure and I used the level where I thought I could just copy a CW signal against the noise. Note that in practice on the HF bands, the band noise will likely be well above the MDS level anyway and so it’s not terribly meaningful. The 20dB S+N/N measure represents a strong, easy to copy signal, so the level this is achieved at is interesting and generally more meaningful I think. This is also objectively measurable through comparison of the received audio level. As the MTR2 does not have AGC in the conventional sense, this does not affect the measurements as long as the signal does not trigger the clipper. In all these measurements, the 20dB S+N/N levels were well below the clipping threshold.

One additional test was to run up the signal level all the way to -17dBm to see what the result would be. Above about S9+16dB, distortion was evident in the earbuds (held well away from the ear by now!), but the signal looked clean on the CRO, so it may well have been the ear buds distorting. Simple S9 signal was uncomfortably loud in my ear buds, but that is a function of the earbuds sensitivity too.

I made the mod by attaching a 10uF SMD tantalum capacitor across pins 1 and 8 of the LM386 opamp. Here are the results running the same tests on the modified radio.

Frequency [kHz] Input level [dBm] Audio [Vrms] Detectable above noise?
7030
Noise only 0.0052
7030
-127 (S0) 0.0076
Yes – Minimum Discernible
7030
-106 (S3.5) 0.0525 Yes, strong – 20dB S+N/N
10116 Noise only 0.0070
10116 -121 (S1) 0.0155
Yes – Minimum Discernible
10116 -105 (S3.5) 0.0399
Yes, strong – 20dB S+N/N
14061 Noise only 0.0049
14061 -126 (S0) 0.0074
Yes – Minimum Discernible
14061 -107 (S3) 0.0488
Yes, strong – 20dB S+N/N

Comparing the before and after values, the MDS seems to improve by about 1 S-point on average (remembering that this is subjective). The big change however, is the 20dB S+N/N measurements. Here there is a 3 S-point improvement on the 20 and 40m bands and a bit more than 2 S-point improvement on the 30m band.  This represents a noticeable improvement in sensitivity.

With the improvement in the sensitivity, the “uncomfortably loud” point comes at about S7 and the noticeable distortion appears at S9 + 6dB in my ear buds. Hence, I think I’ll need to add a volume control to my earbuds!

Overall, this appears to be a worthwhile addition to the MTR2.

Note: I am not experienced at receiver measurements, so am open to advice on improving the methodology of my measurements!

Part 2 (Feb 2015) – Steve Webber recommended mods

Subsequent to the ideas above, Steve Webber KD1JV indicated that the sensitivity could also be improved with a change to the BPF inductors and the gain setting resistor on the 4808 audio amp. I decided to make this mod and see how it compared.

The changes made were:

  • Change L3 and L8 to 10uH
  • Add 22pF across CT2 and CT5
  • Change L5 and L10 to 4.7uH
  • Change R6 to 470k
  • Removed the capacitor I’d previously added across U2 pins 1 and 8.

Here’s the measurements after these changes:

Frequency [kHz] Input level [dBm] Audio [Vrms] Detectable above noise?
7030 Noise only 0.0053
7030
-107 (S3.5) 0.013 Yes – Minimum Discernible
7030
-92 (S6) 0.056 Yes, strong – 20dB S+N/N
10116 Noise only 0.0045
10116 -107 (S3.5) 0.015
Yes – Minimum Discernible
10116 -93 (S6) 0.046
Yes, strong – 20dB S+N/N
14060 Noise only 0.0042
14060 -117 (S1.5) 0.0065
Yes – Minimum Discernible
14060 -96 (S5.5) 0.044
Yes, strong – 20dB S+N/N

Comparing to the original un-modified MTR, there is clearly an improvement. Looking at the 20dB S+N/N figures, the improvement is 4dB on 40m, 1dB on 30m (the only change on this band was the gain setting resistor) and 7dB on 20m (more than 1 S point). These figures are still lower than the modification of adding the bypass cap to the LM386, so I added that back in with a 1uF rather than 10uF (this is based on Steve’s suggestion that the 10uF would increase the gain down too much at low frequency).

Here’s the results with the addition of the 1uF capacitor across pins 1 and 8 of the LM386:

Frequency [kHz] Input level [dBm] Audio [Vrms] Detectable above noise?
7030 Noise only 0.0044
7030
-127 (S0) 0.0065 Yes – Minimum Discernible
7030
-107 (S3) 0.045 Yes, strong – 20dB S+N/N
10116 Noise only 0.0060
10116 -123 (S0.5) 0.0090
Yes – Minimum Discernible
10116 -103 (S4) 0.060
Yes, strong – 20dB S+N/N
14060 Noise only 0.0044
14060 -131 (S0) 0.0060
Yes – Minimum Discernible
14060 -110 (S3) 0.044
Yes, strong – 20dB S+N/N

These results are similar on average to the values for the 10uF bypass capacitor alone with 20m being about 3dB better, 30m is 2dB worse and 40m is 1dB better. The biggest improvement seems to be 20m which is now quite sensitive. In terms of modification effort, it seems that the 20m inductors are worth changing and adding the 1uF capacitor to the LM386 together with the change to the gain setting resistor on the 4808 are also worthwhile. You need to be aware that you may need to add a volume control to the ear buds though!

2 thoughts on “MTR 2 gain mod

  1. I would like to see this same study for the MTR5B. Mine is hissy and I don’t feel it’s very sensitive. At least the audio doesn’t seem as present above the amplifier noise as the ATS3B always was. I added the cap to my LM386, which added more gain (and hiss) to my MTR5B. But I’d like to see that rig improved.

    • Hi Jonathan,

      Yes, I’m sure that would be interesting to do. Unfortunately I didn’t get a MTR5B as I am based outside the USA and hence wasn’t eligible to participate in the release. If I get one in the future, I’d likely do this analysis. Maybe ask on the AT-Sprint mailing list if anyone else can do it?

      73

      David
      VK3IL

Leave a Reply

Your email address will not be published. Required fields are marked *