ANAN-10 TX spurs

FIRMWARE TOPICS ONLY--non-firmware topics will be MOVED
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

ANAN-10 TX spurs

Postby Conrad_PA5Y » Fri Sep 07, 2018 9:23 pm

Hello all.

I have an ANAN-10 that I intend to use on 6m with an SSPA and with 144 and 432 transverters which will be used on the intended band and as drivers for microwave transverters.

There are lot of RX spurs on 28 and 50 MHz but that is the subject of a separate topic.

I will concentrate on in band TX spurs which are at an unacceptable level for my intended application. I suspect that the problem can be solved by better filtering of a yet to be determined SMPSU device.

For all these tests a clean linear supply was used. I investigated the problem as a result of communication with other ANAN-10 users one of which shared plots when running from a battery. Suffice to say that the problem is not from a external supply.

The single tone was generated using the TUNE control in PowerSDR mRX V3.4.9

Here is the first plot at a fairly wide RBW at 8W output.

Image

I decided to investigate further at different power levels and so that I could see deeper I decreased the RBW of the spectrum analyser.

This plot is at 3W out

Image

Now 1W output

Image

Finally from the TX out with the PA disabled.

Image

This is unacceptable for serious VHF use with high power and high gain antennas. Due to the low noise floor of other users at least 90dB suppression of all in band spurs is required. I am quite sure that this can be achieved and that this is just poor filtering of an SMPSU device or ground issues on the 3V3 domain. I would like Apache or the designers to investigate and maybe offer some suggestions as to what modifications may be helpful. I suppose that this is very low priority as it will not help new sales.

Anybody else who may have a suggestion feel free to contribute. I have had a pretty negative experience so far with Apache hardware. I would like to think that this has all been cured on Orion based models but who knows? Anyone with the relevant hardware please check this out.

73

Conrad PA5Y
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-10 TX spurs

Postby w-u-2-o » Sat Sep 08, 2018 1:37 am

The level of performance you measured well exceeds both 47CFR91.307 and ITU-R SM.329-7. Obtaining greater than -60dBc spurious performance in an HF amateur radio transceiver is quite difficult. Are there other amateur radio transceivers that can demonstrate this level of performance or better? If so, is there any data you can cite or show to that effect?
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sat Sep 08, 2018 7:45 am

Of course there are.

The budget IC7300 is better than this. Tomorrow I will do some plots. It is difficult in a superhet but not so difficult from a DAC. It can be solved I am sure of that.

Conrad
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-10 TX spurs

Postby w-u-2-o » Sat Sep 08, 2018 12:56 pm

What spectrum analyzer are you using? From the plots it looks like an HP. I've got an HP8560E here.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sat Sep 08, 2018 2:29 pm

As promised IC-7300 at 100W CW. I had to use a better instrument with an even lower RBW to see what was going on. I used peak hold for 30 seconds.

I think that the standards are woefully inadequate by amateur weak signal standards. That's why suburban amateurs hear loads of noise. Equipment may be compliant but it is not good enough.

The IC-7300 is an inexpensive radio, the ANAN-10 can easily be as good with attention in the right places. Once again I suspect inadequate production testing. I would love to see an Apache internal test report. If you make complex devices then you need to have complex test regimes

73

Conrad PA5Y

Image
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sat Sep 08, 2018 2:33 pm

w-u-2-o wrote:What spectrum analyzer are you using? From the plots it looks like an HP. I've got an HP8560E here.


I used an HP8562A however for the IC-7300 I used an E4406B, I'll re-do the ANAN-10 with the E4406B in a while.

Regards

Conrad
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sat Sep 08, 2018 3:04 pm

Well how embarrassed am I?

The spurs are from the HP-8562A! Here is what it looks like with a proper spectrum analyser, as I expected it to be.

Image

I am not ashamed to admit that I am wrong.

73

Conrad PA5Y
Last edited by Conrad_PA5Y on Sat Sep 08, 2018 3:31 pm, edited 1 time in total.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sat Sep 08, 2018 3:17 pm

That is not 11W but 13W as well.
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-10 TX spurs

Postby w-u-2-o » Sat Sep 08, 2018 4:10 pm

Good to know, Conrad! I hope you post a follow up on the Yahoo Group as well.

You know, I was looking at what were obviously "vintage" HP plots, and something was tickling my subconscious regarding spectrum analyzer spur free dynamic range, but I was thinking more along the lines of noise floors and attenuator settings. I also totally forgot about the spur performance of the equipment from that era. This is a good reminder.

I've found that the Apache hardware I have here, with PowerSDR mRX, is actually better than the 8560E, which is why the 8560E sits on the shelf gathering dust. But there are measurements you can do with even the old HP stuff that you can't with the radio, of course.

73!

Scott
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Sun Sep 09, 2018 12:34 pm

Indeed Scott. In my defense I use modern instruments at work every day and so I had all but forgotten about these limitations. I bought the HP 8562A as I have started to use microwave bands and needed to know what I was transmitting. I had no way of knowing what was happening on 10368 MHz and so it was a necessity. My other analyser the E4406B is a relatively modern FFT based Vector Signal Analyser but the ADC bandwidth is only 10 MHz do not so good for measuring harmonics etc. This was another reason for purchasing the HP8562A which goes to 22 GHz, Basically it seems trustworthy down to -60dBc and could well be out of spec. I have not checked anything other than amplitude accuracy. On the other hand 28.100 MHz may have been an unfortunate choice of frequency. I also checked the ANAN-10 on 50 MHz as well my TS-2000 on 144, 432 and 1296 and IC-746 on 50 and 144 MHz. I happy to say that all of them are better than -70dBc (worst case 1296MHz) and in most cases around -80dBc. This is for CW. I am sure that 2 tone testing will tell another story. I know for instance that the IMD performance of the TS-2000 on 1296 is a no go above 6W.

It is very good to know these things and I am glad that I have the facilities to check.

I am very relieved about the ANAN-10 as it will be the cornerstone of my microwave operations.

73

Conrad PA5Y
f6dro
Posts: 11
Joined: Sun Sep 09, 2018 12:39 pm

Re: ANAN-10 TX spurs

Postby f6dro » Sun Sep 09, 2018 6:23 pm

Hi all ,


I saw this topic development on the forum. Should I conclude that all ANAN100are not the same? Mine is bad , generating 4 (at least) in band spurriis in the 28Mhz band and in all bands in fact.

It is not a spectrum analyser problem as , to be sure , I used an RX to confirm . These products are -60 dB down approx.

I agree it is difficult to get better than -60dBc products on a standard transceiver , because of generating and mixing several frequencies. But there , the spurriis I get are not this case , it is , for the strongest one + and - 385 Khz from the carrier , In my opinion it is not properly filtered PSU in the transceiver.

-60dBc gives quite a strong perturbation for my fellow amateurs up to 100km away and more


The measured spectrum can bee seen here:

https://www.cjoint.com/c/HIjm4wKGnyh


73

Dom
f6dro
Posts: 11
Joined: Sun Sep 09, 2018 12:39 pm

Re: ANAN-10 TX spurs

Postby f6dro » Sun Sep 09, 2018 6:24 pm

ooooops anan10 not anan 100 like mistyped
User avatar
DL2XY
Posts: 116
Joined: Sun Jul 30, 2017 9:47 am

Re: ANAN-10 TX spurs

Postby DL2XY » Wed Sep 12, 2018 6:26 pm

f6dro wrote:ooooops anan10 not anan 100 like mistyped

No problem, they behave exactly the same.


f6dro wrote:... for the strongest one + and - 385 Khz from the carrier...

Maybe you missed the +/- 2.112 MHz ones ?

ANAN100 attached to 2m transverter:
144mhz 1min.jpg
144mhz 1min.jpg (141.48 KiB) Viewed 24508 times


Ok, there are spurs, but they are below legal limit. What would you expect on a pure digital system?
On the other side we have world-class low IMD thanks to adaptive predistortion, nearby perfect bandpass filters and also extreme low phase noise.

Regards Walter
DL2XY
f6dro
Posts: 11
Joined: Sun Sep 09, 2018 12:39 pm

Re: ANAN-10 TX spurs

Postby f6dro » Wed Sep 12, 2018 8:20 pm

Hello Walter ,
in fact I don't care abt 2Mhz spurs ( which are probably coming from the 5V switching power supply).I saw them but was not choked. My problem is in band spurs.
Your spectrogram has been made on 2m , so imagine someone using such a poor tx during a major european contest with 60dBm or more. He calls in the cw band with QRO and is heard in the ssb subband with strong signal for locals .

Good filters , top IMD are software , spurs are hardware.
The other question is : why is Conrad anan10 better than mine?

I'll have a look with my lab on the pcb , maybe I can find the reason?
73
Dom
User avatar
DL2XY
Posts: 116
Joined: Sun Jul 30, 2017 9:47 am

Re: ANAN-10 TX spurs

Postby DL2XY » Wed Sep 12, 2018 11:03 pm

f6dro wrote:The other question is : why is Conrad anan10 better than mine?


I found the Reason: He is using 384kSmp/s !

10m_SR384.jpg
10m_SR384.jpg (136.1 KiB) Viewed 24494 times


But +/- 2MHz are still there.

10m_SR384_5M.jpg
10m_SR384_5M.jpg (132.14 KiB) Viewed 24494 times


73 Walter
DL2XY
f6dro
Posts: 11
Joined: Sun Sep 09, 2018 12:39 pm

Re: ANAN-10 TX spurs

Postby f6dro » Thu Sep 20, 2018 5:48 am

Hi again ,

I got some time yesterday to run more measurements.
You are rigth , spur content depends on sampling. With 384K spurs are -72 (10dB better) , still not enougth and worse than Conrad anan10.

A strong signal on the band is something like -30dBm in the offended receiver , then we get 4 inband spurs at -102dBm , giving respectable jaming signals on the band.
I don't remember wether my anan is a 10 or a 10E , I'm not at home at the moment so I can't verify , maybe it could explain the difference with you and Conrad? Anyway , even -84 db down still jams.
I wonder wether DAC output filter modification could improve the situation ? -100dBc would be nice.
I have seen an SM5BSZ video where he modifies a bit this filter ( for a 100D).
https://www.youtube.com/watch?v=odQBIZi5tAA

Like Leif , I am not afraid at all to put my fingers and my soldering iron into the anan.
I'm surprised the subject does not seem to interest people on the forum.
73
Dom
f6dro
Posts: 11
Joined: Sun Sep 09, 2018 12:39 pm

Re: ANAN-10 TX spurs

Postby f6dro » Thu Sep 20, 2018 7:50 am

Hi ,

me agn ,

I had a look at the 9744 datasheet.
The sfdr explains the spurs we see. Probably my anan is a 10E , I don't know if it is the same ADC in this version , maybe the sfdr is 10dB worse?
Anyway , if it is the case , there is no simple way to improve this. I also had a look at the filter , it is a mini circuit 40Mhz LPF. So it looks like people is going to heam me on several frequencies on the band. I am lucky I use the anan on microwaves , on 2mtr , it is a serious pb.
73
Dom
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: ANAN-10 TX spurs

Postby Conrad_PA5Y » Thu Oct 04, 2018 12:13 pm

Hello Dom sorry I missed your subsequent posts. I checked mine again and it really does appear to be better. I am beginning to doubt my instruments again! I will take it to work and check it on a more modern calibrated R&S spectrum analyser.

Nobody is interested because most of them have such high noise floors they can't hear anything.

73

Conrad PA5Y

Return to “Hermes Firmware (10 & 100)”