"Sticking" Relay on ADC 1 and 2 Receive

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

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Fri May 11, 2018 7:30 pm

Hello Scott I have an excellent reason for not wanting to do that. There IS the possibility that ANT2 can shove some tens of watts up the rear end of my LNA. My only defense against this calamity is the database xml file and I just don't trust either it or myself. IF I should make a mistake (which I often do) it means performing surgery on something that is high in the air. This is not trivial.

Of course it might well be possible to isolate the LNA with an external relay, this is rather easier than performing surgery on the ANAN-100D but it still leaves a bad taste. When I see sockets labelled Ext1 and Ext2 I expect them to work as they would on any other transceiver. I had no idea that there was such a spread in performance on these units. Some people see these as 'niggles' I feel that they are unforgivable.

I may well have to use ANT3 as a transverter input as well, so that is another relay. I can see that I will be making an interface box.

Regards

Conrad PA5Y
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Fri May 11, 2018 9:33 pm

'Disclaimer: I do not work for Apache, nor am I compensated by Apache. I did volunteer to help run this forum.

Conrad_PA5Y wrote:I had no idea that there was such a spread in performance on these units.
There is no dramatic "spread in performance". There are a very few people who occasionally have radios that are not working right, either out of the box or broken later on. No manufacturer can turn out a 100% perfect and 100% reliable product, not Icom, not Kenwood, not anyone. Not at any price point, either, whether it's a $1000 radio or a $4000 radio. And this is even more difficult for a small company like Apache that turns out comparatively few units a month. Such a situation simply does not lend itself to the ability to qualify or validate a design such that it is reliable to the same extent as a design that is being produced in the 100's or 1000's of units a month. There is just not enough volume to make it economically feasible. When you buy low volume, specialty consumer electronic hardware, one has to expect it is going to be less reliable. The manufacturer of such a device generally compensates for this with a very effective and generous repair process. Everyone I know who has engaged with Apache on a repair has been very satisfied, including myself.

And you must remember that Apache supplies the hardware ONLY. The software and firmware come from an open source community that is not comprised of professional developers, does not subscribe to industry best practices or processes for development, testing, validation, documentation or any other dozen factors associated with software development that you might care to name. As a result, this software and firmware is NOT going to be 100% reliable, stable and bug free.

Some people see these as 'niggles' I feel that they are unforgivable.
It's not so much that they are niggles but rather that they are offset by so many other positive or unique things that the "niggles" are just considered the cost of doing business. Remember that 90% of the radio is PowerSDR, the hardware unit is nothing more than some filters, switches, amp, ADC, FPGA, DAC and glue. It doesn't seem like a lot for the money, but they are very low volume so there is no economy of scale. So you are paying extra and you have "niggles". BUT...if you want:

1. The ONLY radio that provides adaptive predistortion transmit linearization
2. The ONLY radio that provides rack/DAW type audio processing built in
3. The ONLY radio that provides NR2 noise reduction
4. The ONLY radio that provides coherent beamforming (aka diversity)
5. The ONLY radio that provides at least a half dozen other obscure features such as true independent sideband AM receive
6. Arguably the best DSP filtering in the business (brick wall, low latency, etc.)
7. AND at the same time hangs with the top 10 or 20 on the Sherwood list depending on which hardware unit you pick

well then some things are going to have to be traded off. Quality and reliability might not be six sigma. The code might not be bug free. The documentation might not very complete. And so on. But if you don't need the unique features of PowerSDR, and would rather have something that is 100% super reliable both out of the box and ten years down the road, with effectively zero software/firmware bugs, and a big, thick manual, then, respectfully, you might be happier with another radio, and I'm serious, I'm not saying this to be sarcastic. I'm obviously a huge evangelist for Apache and PowerSDR, but I still point a lot of people at the 7300 or Flex, just because I know they will be much happier.

When I was looking to upgrade from the Flex 3000 I had I agonized for at least a month over whether it should be this crazy Apache Labs open source thing, or if it should be the more mainstream Flex 6300. Obviously I went with Apache and PowerSDR. It's been a bumpy road for sure, but I'm certain it's been a lot more fun and interesting, at least for me, because I'm interested in being on the cutting edge of technology. Not everyone is happy or comfortable hanging it out over the edge like that.

One last point: Apache has reached a level in the market where it appears to be a mainstream choice. Heck, its products are distributed by Gigaparts and HRO, right? But it's really still a very niche choice. Except in one, very, very important regard: reliability and customer satisfaction. I'm pretty certain that if things were a train wreck in those departments neither Gigaparts nor HRO would carry the Apache brand, as the costs of processing both repairs and returns are quite high.

73!

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

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Fri May 11, 2018 11:10 pm

Well all I learned from that lengthy tome Scott is that you are an ANAN evangelist. This was not a surprise to me or anyone else I suppose. The 'message' was obviously intended for bigger audience than me, that is clear - and understandable.

However some of it is guesswork. I'll let you work out which parts.

I shall now deliver a novella of my own.

I found this statement to be quite preposterous.

It's not so much that they are niggles but rather that they are offset by so many other positive or unique things that the "niggles" are just considered the cost of doing business.


Considered by whom? Not me that's for sure.

I expect a radio to work as described, it doesn't and therefore I am not very happy about it. This is a perfectly reasonable position to take. I had already considered all of the plus points that you made in your post and indeed that is why I chose to buy the ANAN-100D. I was very attracted to PS because I have a conscience about the signal I radiate, it is particularly important on VHF in an area of high station density like central Europe. I have some very specific requirements that had the radio worked as expected it would have been perfect for my applications. However it has proven to be quite unreliable.

I get the feeling that you think I am in some way incompetent. I am not, in fact quite the contrary. As I repeatedly have said the PowerSDR/ANAN combination is a means to an end for me. Yes it is very interesting and opens up some new possibilities. However I bought it so that I could work DX on 6m during harsh conditions. I do not want to spend too much time tinkering . I have obviously made a mistake, one that I will have to live with. There is no way that I can sell this radio to another amateur knowing what I do. So I will have to find a way to make it work - but it will have to wait until the DX reduces significantly.

You (we) are NOT on the cutting edge of technology, we are using modern techniques but cutting edge is pure hyperbole, so maybe drop that from the sermon? Also maybe reduce the trendy buzzwords 6 Sigma, good grief! You are not at work. I am prepared to bet that you have or had a background in marketing, If not you have been spending far too much time with marketing people.

I have a legitimate reason to complain and no amount of preaching will stop me.

You did not comment on why I choose not to use the ANT2 input for my LNA RX feed which was the subject of my last message. Instead of discussing the technical issue you delivered a sermon.

By the way I do not wish Apache or anyone else involved in the project any harm but I have the right to be disappointed and complain at a lack of basic functionality . I do not see any of the negatives or 'bumpy road' comments listed prominently. This is not meant to be a commercial venture so where's the up front honesty? Having it buried somewhere in this one man forum does not go far enough. The hardware/software combination should work as expected. Reliable TX/RX changeover is a pretty basic requirement.

You can't have it both ways.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Sat May 12, 2018 12:21 am

Conrad, I wish you the best of luck in your endeavors. I hope you send your radio for repair so that you can either put it to work or sell it with a clear conscience. Or perhaps sell it as is at a deep discount.

73,

Scott
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby K9RX » Sat May 12, 2018 1:54 pm

Conrad and Scott....

I've been on the side holding back, as Scott well knows that is not my normal tack. I'm quite vocal with "opinions" (generally those are based on present perceived facts)... but the one thing that set this situation aside is is this a fault in a particular unit or is this a design issue.

If this is a fault then I fully agree that we have to be more tolerant of that condition. Although as a designer and manufacturer of a product for a niche market, http://www.ServoCAT.com, I can say that it absolutely can be designed to be nearly 100% reliable. I have nearly 2000 of my main product in the field starting in late 2001 and have extremely few problems with it - I KNEW being a one man company that if i had ANY reliability issues it would quickly consume my time - and time is my most valuable asset. I designed it to be bulletproof UP FRONT. So it CAN be done. But again - if it is a hardware fault that we have to occasionally put up with - even my FTdx5000 developed several issues. And my first 8000DLE failed completely - I had to get a replacement (and that was wonderfully handled by the way, thank you Doug).

But if this is a design issue, which it is on the 7K/8K, then Conrad has every right to be upset. One doesn't pluck down thousands of $$ and HOPE that everything works as advertised. As you indicate it is going mainstream and yet there are items that should just work that don't, like the sticking relay problem (on 7K/8K, Scott has said that the hardware is different on a 100 - although your earlier response to this thread says the 100 is the same!). For the guy to come on the other group and say "stop your complaining - after all you have a back up radio" ... GEEZ! SERIOUSLY? If you buy a high end car and it doesn't do something basic are you supposed to feel comforted that you have a 2002 Ford truck sitting next to it? It isn't asking anything additional that ALL back panel functions should work - period.

I don't believe this onus is on the volunteers - this is a radio related issue (read: hardware) and as such Apache Labs needs to not only solve it but be proactive. TOO much "oh this is volunteer" goes on - and Apache Labs benefits off this - they're let off the hook as if all they have to do is provide some bit of hardware and they're done until the next release possibly years later. We as a group need to push AL to test better - ALL functionality. And retest, a complete battery of tests, every time there is a proposed new software release. It is after all their part of the equation. Letting them off the hook after having paid them ALL the money that is paid out - none to the software developers - this should be baked in to the cake. It should be absolutely required. (I'm still waiting on the replacement back panel for the poorly designed RCA flex issue)

There are MANY guys that only use the radio for ESSB type use or phone in general ... my guess is the majority only have ever used 20% of the function/features. My gut says there are far fewer that use it for serious work like I believe I do and like it is obvious Conrad does. This is not to suggest one is better or more important. But it is to suggest that problems that are seen by those that use most all function are often not seen by the majority. That doesn't mean that, the "majority", rule! It doesn't mean that we should be pooh-poohed. Patronized. I need it to be at least as good as the best high end radio out there - it isn't being sold as a "tinkers radio" ... it is being sold as a mainstream radio with high end features (and for the 8K at a premium price!). And it won't be taken seriously by DXers, contesters, weak signal guys, those that push the envelope unless and until AL steps up to the plate.

I had a Flex 6700 in here for 4 months doing A/B RX testing comparing it to my then primary radio, a Yaesu FTdx5K. Never not once during that time did I find a signal, weak CW or crowded band phone that I couldn't make as intelligible on one compared to the other - either way. So it was no better than the 5K (both being excellent) ... on top of that their platform, although "pretty" and better designed in many aspects, was too static. The last straw for me on that radio was when I finally went to transmitting on it, where the processor is critical (after all if done right it is free "gain" on TX) I found my Heil mic/headset wouldn't interface to it. The audio was too low. I was told "get an external preamp" ... that was it. A $7K radio that doesn't interface to one of the most ubiquitous headsets out there. I sold it. The point is these things should just work - period. One shouldn't have to add lots of external clutter to use what is there on the box (and most would have a difficult time technically doing so at least on some issues) or feel denigrated for pointing out the lack of function.

I love the radio ... and more importantly I love the software. It is without a doubt one of if not the best out there. I am still here because of all the features. Like Conrad one of the primary reasons why I bought in is cleanliness ... on phone that means Puresignal - I run BIG antennas and legal power - I absolutely want to be a good steward of the bands ... add to that all the features and functions and it is a tremendous platform. But lets demand the manufacturer play more of a role here... and stop suggesting that those expecting things to work "might be better off with another radio". As can be said: "work with me (us) here".

Gary
K9RX
User avatar
Tony EI7BMB
Posts: 651
Joined: Sun Apr 09, 2017 2:31 pm
Location: Dublin
Contact:

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Tony EI7BMB » Sat May 12, 2018 4:00 pm

Well put Gary, I agree with all your points especially that Apache should take more of a lead here. I have nothing but admiration and appreciation for the work the volunteers do.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Thu Jun 14, 2018 11:19 am

I have to say that I am very disappointed by the lack of response from Apache Labs. I have had to use ANT1, 2 and 3 for all my RX requirements and protect transverter IF RX outputs with additional coax relays. The configuration of transverter switching using the DIO on the 25 pin connector also seems to be unreliable. This is a real shame.

The performance of the RX/TX itself is excellent. I have a truly state of the art 2m system using a transverter but I had to jump through hoops to get there.

This RX/TX change over reliability issue on the Ext1, Ext2 and Transverter RX sockets is a travesty and has ruined the whole ANAN experience for me. It took me some weeks to feel able to circumvent these failings.

Regards

Conrad PA5Y
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Thu Jun 14, 2018 11:57 am

Have you contacted the factory directly?
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Fri Jun 15, 2018 9:09 am

No because I do not wish to ship the unit to India with the associated risks, nor incur the associated costs. I have workarounds so I will live with it. However I reserve the right to complain periodically.

Apache MUST be aware of this problem and are not interested in proactively solving problems for 'legacy' products such as the 100D. I have made my point repeatedly, this stuff SHOULD work. The only reason that it is not more prevalent is that most users never use these sockets. There is a problem and I am not alone.

As for the ANAN control.......

Don't you use some 3rd party software that takes band change info to configure your switching? I am going to use good old fashioned switches

Regards

Conrad PA5Y
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Fri Jun 15, 2018 12:52 pm

For external control I use both the open drain outputs on the accessory connector and also DDUtil with an Automation Direct "Click" series PLC. Both work fine. The DDUtil stuff is extremely powerful but not well documented but I can help with some of that. The open drain outputs will change state at odd times, like when switching transmit profiles, but go right back to the proper configuration for the band. For that reason I recommend their use only for band switching applications, not for control of things that want to remain in the same state while tuning around.

If you would like to start a discussion about this I'd recommend starting a new topic in this same subforum.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Sat Jun 16, 2018 8:06 am

Hello Scott, I will do so when it is pertinent. At the moment I have more fundamental stuff to do such as 10MHz distribution, a rack box for the LO sources and transverters to put in boxes. It feels like a mountain, but eventually I will be ready to do this. Actually I will only require band changing to route a 144MHz transverters TX out and RX inputs to other uwave transverters. Sounds as though I will be fine.

73

Conrad PA5Y
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Sat Jun 16, 2018 9:11 am

Arggghhhhhh, I also have the same sticky relay issues with the ANT1, 2 and 3 inputs, just much less frequently. I think that the unit has to go to Apache labs after all.

Conrad
User avatar
Tony EI7BMB
Posts: 651
Joined: Sun Apr 09, 2017 2:31 pm
Location: Dublin
Contact:

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Tony EI7BMB » Sat Jun 16, 2018 10:17 am

Conrad you may remember my 6M LNA failed and even though I bought direct Abhi at Anan was able to arrange for it to be looked at in the UK and I think there may also be a tech in Spain, perhaps this might be an easier option than sending to India or USA.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Sun Jun 17, 2018 7:30 am

Hello Tony, I have contacted apache support. I expect that they will advise me what to do on Monday. The radio is unusable as it is so I may as well get this sorted.

Regards

Conrad
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Sun Jun 17, 2018 11:02 am

If a large number of relays are "sticking" it sounds like it could be a problem with one or more relay control ICs. If that turns out to be the case then that would be a different problem than the occasional balky relay that some people have experienced and not a fundamental design flaw, but rather an unfortunate hardware failure.

I have no doubt Apache will make it right.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Mon Jun 18, 2018 8:56 am

ICs used as drivers do not usually fail intermittently but fail totally, anyway I would be highly delighted if you are right Scott. Hardware failure is preferable to hardware design errors.

No response from Apache yet. I expected some response this morning but I'll give it another day before I start moaning :-)

I did ask them if there are any tests that I could perform to narrow things down before I ship it off for repair.

Regards

Conrad
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Mon Jun 18, 2018 10:19 am

If you take a look at the schematics (links available in the schematics sub-forum) you could probe the boards to see if the relays are getting the right control signals at the right voltages.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Mon Jun 18, 2018 11:20 am

If Apache do not have any specific suggestions I will do exactly that.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Wed Jun 20, 2018 11:16 am

Apache did contact me as did Doug W5WC. They suggested trying a few combinations of FW with PowerSDR mRX PS v3.4.9, starting with Angelia_r6.0.rbf.

The good news is that this appears to have fixed the problem. I have not run exhaustive tests yet but will do tonight. I have tried 1000 transitions on EXT1 and EXT2 inputs without a single occurrence of the problem. Tonight I will run 10000 transitions on EXT1, EXT2, ANT1, ANT2, ANT3 and TVTR IN.

I am quite confident that the more extensive tests will be sucessful to my considerable relief. It is a difficult situation as it is only 'some' hardware/FW combinations that exhibit the problem. Until the underlying cause is understood this is difficult for the hardware and software teams to manage.

I would like to thank Apache Labs, Scott and Doug for their invaluable help.

Regards

Conrad PA5Y
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Wed Jun 20, 2018 8:16 pm

I have done 3 inputs 10,000 transitions and the sticky relay issue still exists but at a greatly reduced regularity. The occurrences are now approximately 2% of the time. A factor of more than 10 times less often. This is tolerable to me.

I am not sure if the reliability is affected by the PTT method. I will try with an external PTT and a via CAT command and report back. The truth is that I now have a manageable problem so I can use the ANAN in my system.

73

Conrad PA5Y
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby w-u-2-o » Thu Jun 21, 2018 8:10 pm

I know you are probably not going to do this, but it would be so very useful to know if the problem is because the relay driver register did not get written correctly or if the problem is really a balky relay. My guess is the former, since the new firmware substantially improved the problem.
User avatar
Conrad_PA5Y
Posts: 101
Joined: Fri Jan 19, 2018 8:11 pm

Re: "Sticking" Relay on ADC 1 and 2 Receive

Postby Conrad_PA5Y » Thu Jun 21, 2018 8:49 pm

Hello Scott. It must be a register or timing problem as a ten fold reduction in the frequency of the events cannot be anything else. The relay did not suddenly become 10 times better.

Anyway tonight I did a pure signal mod pretty much exactly as you did and something dawned on me after doing so. The ADC input that bypasses the HPF board does not suffer from this problem at all. As I use an external LNA and very high quality band pass filter on 6m there is absolutely no reason why I should not use this direct connection on RX for 6m and for PS on TX. I have a -50dB Xtronic coupler and it works superbly for PS now.

I am also going to use this self same input for transverters, use the PTT for PS on 6m and a 4-way sma switch to switch the RX path between 6m and the various transverters. This makes me totally immune to the problem.

So this is what I will do.

Regards

Conrad

Return to “Everything Else: Antennas, Relays, Switches, Power, Grounding, Cooling, etc.”