apache-labs.com

Concerning the voltmeter reading in the lower left on Thetis..... Mine constantly fluctuates between 13.1 and 14.3 volts. It's never steady.
Same with 3 different power supplies and other voltmeters indicate no variations.
What causes this?

Not having any more information than you gave us, there are a number of variables that could be contributing.

1) Where are you connecting the other voltmeters you're referencing? Directly at the power supply? That's not where your radio is connected (it's connected at the other end of the power cord).

2) How long is the power cord that connects the power supply to the radio? What gauge wire is the power cord?

3) What model radio are you using?

4) Under what operating conditions are you seeing the voltage swings? Is it only during 100 watt digital mode transmission (I hope not!)?

5) What other devices do you have connected to the same mains circuit as your power supply?

6) What is the overall load on the transformer feeding your house? How many other homes are connected to that same transformer?

My 7000 dle mk2 does this also. It happens in rx mode. I never checked it while transmitting. I can't remember the exact numbers but the voltage bounces around +/- .2 volts. I use a linear Astron 50 and the voltage is rock steady. It's annoying so I just turned off the volt/amp meter.

My 7kdle mk2 also varies approx. +/-.2 volts but the OP is seeing a swing of over a full volt.

K1LSB wrote:My 7kdle mk2 also varies approx. +/-.2 volts but the OP is seeing a swing of over a full volt.

Thanks Mark. I had my head in the clouds.

I fired up the anan to check my memory. It is fluctuating between 13.3 and 14.0 volts, and 0.0 to .2 amps at idle. On TX at 100 watts CW I have 12.8 to 13.1 volts and 14.9 to 15.0 amps. I didn't crawl behind the desk and probe into the power cable but I will if that will help anyone.

1. Power cord is 3 feet long 8 ga. stranded.
2. 7000 DLE mk 2
3. Condition is in receive mode, AF gain at zero.
4. The variations are greatest during receive mode. During 100w CW key down I see 13 to 13.2 volts with occasional variations in between.
5. No other devices are connected.
6. Home has 400 amp entrance. Astron 35 is on a 30 amp 120v line.

I don't think the above matters because Voltage at the PSU is steady, 13.8vdc measured with digital meter.

n4lq wrote:1. Power cord is 3 feet long 8 ga. stranded.
2. 7000 DLE mk 2
3. Condition is in receive mode, AF gain at zero.
4. The variations are greatest during receive mode. During 100w CW key down I see 13 to 13.2 volts with occasional variations in between.
5. No other devices are connected.
6. Home has 400 amp entrance. Astron 35 is on a 30 amp 120v line.

I don't think the above matters because Voltage at the PSU is steady, 13.8vdc measured with digital meter.

I have the same issue. There is a great voltage fluctuation in receive, but not in transmit. It doesn't bother me one way or the other. If it will help others, here is my setup:

1. Nine month old 7000DLE MKII
2, Power cord is 3 foot long 8 ga. stranded
3. When receiving the voltage fluctuates from 13.1 to 14.1 volts dc
4. When going into transmit the voltage momentarily drops back to 13.0 vdc. It immediately recovers to 13.8 volts.
5. Speaking during transmit SSB results in the voltage fluctuating from 13,8 to 14.0 volts.
6. When transmitting two tones it never moves from 13.8 volts.
7. The only other device connected is a digital voltmeter connected at the output of the PS. The voltage there stays at 13,8 vdc whether receiving
under full load transmit. It does momentarily drop back to 13.7 volts when radio is first put into transmit. It instantly recovers to 13.8 vdc
8. I'm running Thetis v2.8.11 and Protocol_2_v2.0b firmware.
9. I have tried 4 different firmware versions lately and it makes no difference which one I'm on. I was trying other firmware versions chasing
another issue. I used each one for a period of time and I did pay attention to the fluctuations even then.
10. I noticed this also in earlier versions of Thetis

Like I indicated earlier, this doesn't bother me. I just thought it was interesting. Maybe this will help others.

Please Try This:

Setup / Display / General CPU Meter (ms)

Change to 9999

Mine is more stable but simply because the refresh takes so long.

Sounds like the voltage measurement needs some additional averaging before being displayed for sure.
/Doug

My Voltmeter on receive is all over the place. Never stops bouncing displaying varying voltage between 11 to 18v. Do I have a hardware problem? Been like this since I received my black 7000 a few weeks ago.

I don't have a 7000, but if you look at the following Orion MKII schematic snippets you can see the 12VDC is sent to a voltage divider, the output of which is limited by a zener diode, then sent to an 8 channel ADC.

You'd have to put a scope on the 5V analog voltage reference input to the ADC to see how well behaved it was. Similarly, you'd have to put a scope on the ADC input in question to see how noisy it is. After that it becomes a question of how well the firmware does at reading the ADC and how much filtering goes on in the software.

IMHO it's not super important stuff, but it is certainly understandable why it is annoying that it doesn't work well.

Thetis code... protocol 2

//Bytes 53,54 User ADC0 [15:0]
prn->user_adc0 = prn->ReadBufp[53] << 8 | prn->ReadBufp[54]; // AIN3

53,54 are actually 57,58 in the port 1025 high priority status packet (page 46 of the protocol manual)

Thetis takes 100 of the adc samples, and averages them



EDIT, just to add this code is not used any more, which I note uses the 4.7k and 820ohm as per the schematic you included Scott.

//Bytes 45,46 Supply Volts [15:0]
prn->supply_volts = prn->ReadBufp[45] << 8 | prn->ReadBufp[46]; // 49 + 50 in the 1025 packet



So, I guess we need to look at the schematic for AIN3 ?

Richie.

It's a little bit of chalk and cheese going on here, Richie, I think.

AIN3 is the 12VDC power measured at the input to the PA, on the PA board. It's measured after it comes out of the ACS713 current measurement IC. The 12VDC is processed through an LM358. I haven't checked the gain of that circuit vs. the code yet, but I suspect it matches. The output of that op amp is taken off the PA board through that Berg connector, goes over to the rear panel breakout board and then is taken into the Orion MKII board via the J16 connector. I hesitate to post too many schematic snippets since Apache has asked for the schematics to not be publicly posted, but the snippet that shows the sensing circuits on the PA board is shown below. I'm sure you have them anyway, Richie.

AIN4 is the amp current measured by the ACS713 and processed through other half of the LM358. Same signal routing as above.

AIN6 is the 12VDC as measured on the Orion MKII board, as previously discussed.

Are they all supported in the P2 interface on the high priority packet? They could all be displayed on the status bar, theoretically. I do have the amp voltage and current being displayed on the 8000. I have not checked to see if the scaling factors are different in the code or handled in the hardware on the 8000 PA board, the major difference being that the 8000 amp power supply voltage is 50V. The 50V display I see here is well behaved.

yes, but the take away from the code is that User ADC0 is being used to display the PA voltage (commented as AIN3 in the code) and uses bytes 57+58 from the HPSP. Bytes 49+50 which are defined as Supply Volts (whatever AIN that is I do not know as it is not commented in the code, probably 6?) is not used in Thetis right now.

I am sure Doug knows what is going on with this stuff. Wasn't there some issue with the current read out on the newer 7000's reported as well at some point? I notice there is a change based on radio type for the calculations. (computeMKIIPAAmps is the function if you want to look in source)

Richie.

grabbing some data, P2 7000dle, and playing around with some different smoothing. Sampled every 10ms.

I have changed the way the adc's are sampled in 21k9c which should be out over the weekend some time.

Previously, a thread would sample PA volts for 100 iterations, with 1ms delay between each read. After that, it would then read amps for 100 iterations, with 1ms delay between each read. Then it would wait an additional 600ms and repeat the process. Consequently, we would have blocks of reads that were close together spaced around 800ms as the total process would take ~800ms to complete.

I have changed this, to take a single volt and amp reading, every 8ms. So in 800ms we would have taken around 100 samples of each, the same total samples in the same time as above, but now spread out over the duration. These are dumped into a fifo queue as they come in, and then averaged when a volt/amp status bar update is needed.

Perhaps this massive volt-swing that people have noted is related to more recent 7000s, who knows, but the change above should eliminate any step changes related to sampling at nearly once per second.

Richie.

Hello Richie,

would it be possible to activate the voltage and ampere display for all radios if you want to, like on the Anan 7000 or does this function already exist?

René

dj1yr wrote:Hello Richie,

would it be possible to activate the voltage and ampere display for all radios if you want to, like on the Anan 7000 or does this function already exist?

René

possibly, I don't know what is returned to be honest through the various firmware for all other radios

Richie.

ramdor wrote:yes, but the take away from the code is that User ADC0 is being used to display the PA voltage (commented as AIN3 in the code) and uses bytes 57+58 from the HPSP. Bytes 49+50 which are defined as Supply Volts (whatever AIN that is I do not know as it is not commented in the code, probably 6?) is not used in Thetis right now.

I would say it's a near 100% bet that Bytes 49+50 is AIN6, I'd call that supply voltage, whereas Bytes 57+58 is AIN3, PA voltage. I'm assuming you have also identified the bytes for AIN4, PA current?

ramdor wrote:grabbing some data, P2 7000dle, and playing around with some different smoothing. Sampled every 10ms.

Your prototype smoothing algorithm looks much better. A LPF and/or rolling average approach as opposed to chunking it up will provide not only a more stable answer but the opportunity for a higher/smoother update rate on the UI.

ramdor wrote:possibly, I don't know what is returned to be honest through the various firmware for all other radios

AIN6, supply voltage, is available on all ANAN models from the bare board Hermes on up to the latest 7000.

AIN3 and AIN4 are only available on the 8000 and 7000 variants.

It would be nice to see another shot be taken at the presentation of this data. Perhaps previous efforts might have been abandoned because of a sub-optimal approach to data smoothing.

Supply voltage is perhaps most important. There have been instances in the past where melted fuse holders and partially melted fuses have caused issues, and also on 6M where current requirements are high and outstripped the available power supply, all of these things contributing to voltage drop during transmit.

Amp voltage and current are just nice-to-haves. To date there's never been a published case where someone with an 8000 (where this is working) said "My amp current is way too high/low."

I have the 7000 MK2, received in May 22. Running Thetis 2.9.0 My voltage goes from 13,6v to 15v, I have tried 3 linear PSUs, I even bought a Samlex switching PSU and the flucuation remains.

I'm going to try this tnx to N4LQ

Setup / Display / General CPU Meter (ms)

Change to 9999

I'll report back, on the other hand I love this radio for RX and TX.

73's from Belgium.

On my 7000DLE Mk1, P2 v2.1.18, Thetis 2.9.0, it fluctuates from 14.1vdc - 14.2vdc during RX. Power supply is rock steady. Voltage at the back of the radio is rock steady. I have a bout 8 feet of 10ga wire to the PS, which is a home-brew 100a switching supply. On transmit, it only fluctuates 1 tenth of a volt.

So not seeing what some of you guys are.

Perry

You can't compare with voltmeters which are all horribly slow to respond to voltage fluctuations.....

I'm working on a patch to smooth this out...a bit more averaging is needed.

I re-wrote the volts/amps averaging code a while ago. Every 8ms a volt and current reading are taken and are stored in a queue of capacity 100 for each. At the display interval defined by cpu meter interval (yes I know), these queues are averaged, converted and displayed as V and A. As we are rounding to display the text, a tiny change may cause the flicker between 13.8 and 13.9v for example. Increase the queue size too far and it will take ages to settle to the required value.

I have added some logic that will only let the display change if the deviation between previous readings is > 0.05, which should prevent a flicker between 13.9 and 13.8 if just on the edge case. I have also increased the queue size for volts.

Also to note, there is a change in my new version (released soon(tm)) to adjust the current calculations, and I will be looking to do the same for volts to hopefully get around the issues with new hardware somewhat.