|
OK... I destroyed C202 and had to replace it with one on hand, but I was able to attach it vertically so that it made contact with the voltage doubler side alone, and with a tiny wire, I connected its input to the RF output of the output transformer.? ?So the HV supply should now be working.? ?(I did not test that but all the connections were holding via solder) I then carried out the exact same ttests with commercial pads and attenuators and measured the output.? ?It is within a db or two of what I measured with just the FILTER in line.? ?So there was indeed a TINY bit of potential degradation, but it sure isn't much.? ? The horrible harmonics are NOT coming from the HV voltage doubler at least not on 15 meters....? ?That is very good news.? ?See the photos below.   This leaves I think three main other possibilities a)? production by the input switching diode b) leakage through other filters or around somehow c)? production by the output switching diode Gordon KX4Z
|
For reference:? ?15 meter drive is at "84" and previously this would have produced a shade under 15W output.
Gordon KX4Z
|
Gordon,
Good work so far.
A subsequent test could be pulling the 1n4007 diodes for all but the one filter under test.? That would determine if the blow-by is coming through the bias circuit or the diodes themselves causing the harmonics.? If that is not the issue, adding back the other diodes one filter at a time might show some interaction.
Just suggestions.? I would help if I had v2.
73
Evan
AC9TU
|
Thanks very much for that suggestion, Evan!
toggle quoted message
Show quoted text
On Nov 2, 2023, at 19:21, Evan Hand <elhandjr@...> wrote:
?Gordon,
Good work so far.
A subsequent test could be pulling the 1n4007 diodes for all but the one filter under test.? That would determine if the blow-by is coming through the bias circuit or the diodes themselves causing the harmonics.? If that is not the issue, adding back the other diodes one filter at a time might show some interaction.
Just suggestions.? I would help if I had v2.
73
Evan
AC9TU
|
Articles with similar circuits:
|
|
|
Yes, I agree. ? But I don’t know where it comes from. ?
toggle quoted message
Show quoted text
On Nov 2, 2023, at 23:40, Ashhar Farhan <farhanbox@...> wrote:
? Gordon, I noticed that there is a lot of noise around the signals fundamental. That shouldn't be there either. - f Articles with similar circuits:
|
NEW EXPERIMENTS TONIGHT 1.? Now the "input switching diodes"? sides of the low pass filters were reconnected, in addition to the HV voltage power supply, and the ISOLATED highest low pass filtter is still used to directly take the output of the amplifer and go to the output. 2.? See photo of the SCHEMATIC, the photo of the soldering, and the photo of the output spectrum 3.? The 2nd harmonic is now about -42.5dBc,, the 3rd harmonic is still way way down.? ?Having the INPUT diodes connected -- but no connection from the OUTPUTS to anywhere, didn't make a terribly large bit of difference.    73 Gordon KX4Z
|
Very Differrent Result -- OUTPUT DIODES from other filters also now connected. 1.? The 21 MHz signal still goes directly through the upper LPF to the antenna 2.? The HV voltage supply is connected (but I have not measured whether it is working properly and this is crucial!) 3.? The input diodes fromthe OTHER LPFs are now connected 4.? And new here, the OUTPUT diodes from the OTHER LPFs are now additionally connected. AND WE HAVE HARMONICS --  Here is the schematic that I think applies:  And here is a photograph of the actual soldering:  SO NOW there are lots of questions: 1.? Is the HV power supply working? 2.? Are the switching diodes all reverse biased (every one of those should be reverse biased? 3.? How does this happen, since all of the other filters' cutoff frequencies are WAY LOWER ANYWAY!!? Even if one of the other filters were TURNED ON, it should not be passing 42 MHz nor? 63MHz. 4.? Are the back-biased diodes somehow creating the harmonics??? Anyone got ideas or explanations???? ?I'm all ears as I work toward the next experiment, whatever it is. 73 Gordon KX4Z
|
Gordon,
Can you remove the long jumper runs for this last test and just jump across D15 and D19?? This would be to eliminate any bypassing of the filters through the jumpers.
73
Evan
AC9TU
|
Evan, I get what you are going for— that it might be coupling between the loops, however take a look at this post
Where I used to very short jumper across first the input and then the output diodes individually I have the highest low pass filter, with all of the other still connected which is very similar to what you were asking for,…., and it made single-digit changes with still very high harmonics.
I think that suggests that it really does have something to do with all the other systems being connected, possibly their diodes
An easier change that I can do is to simply remove the tack solder point on the output diode (leaving all of the loops of wire still physically present ) and check again to see what spectrum look like. I bet even with the wires still in place but ***no longer soldered ?to the output diodes are the other filters***, it will be again far cleaner
I’ll try to get that done early tomorrow.
Thanks for your review! ?We are closing in on it
73? Gordon
toggle quoted message
Show quoted text
On Nov 9, 2023, at 21:52, Evan Hand <elhandjr@...> wrote:
?Gordon,
Can you remove the long jumper runs for this last test and just jump across D15 and D19?? This would be to eliminate any bypassing of the filters through the jumpers.
73
Evan
AC9TU
|
Here I have disconnected the connection to the output diodes but left the loop of wire within a millimeter or two of the solder point so it could pick up capacitively
The output is still pretty clean minus the connection to the output diodes
So I would suggest it has nothing to do with the length or positioning of the wiring. ?
Gordon Kx4z?
toggle quoted message
Show quoted text
On Nov 9, 2023, at 22:18, Gordon Gibby via groups.io <docvacuumtubes@...> wrote:
? Evan, I get what you are going for— that it might be coupling between the loops, however take a look at this post
Where I used to very short jumper across first the input and then the output diodes individually I have the highest low pass filter, with all of the other still connected which is very similar to what you were asking for,…., and it made single-digit changes with still very high harmonics.
I think that suggests that it really does have something to do with all the other systems being connected, possibly their diodes
An easier change that I can do is to simply remove the tack solder point on the output diode (leaving all of the loops of wire still physically present ) and check again to see what spectrum look like. I bet even with the wires still in place but ***no longer soldered ?to the output diodes are the other filters***, it will be again far cleaner
I’ll try to get that done early tomorrow.
Thanks for your review! ?We are closing in on it
73? Gordon
On Nov 9, 2023, at 21:52, Evan Hand <elhandjr@...> wrote:
?Gordon,
Can you remove the long jumper runs for this last test and just jump across D15 and D19?? This would be to eliminate any bypassing of the filters through the jumpers.
73
Evan
AC9TU
|
Here is my first measurement of the voltage produced by the RFHV voltage doubler. ?
I’m measuring at C201 with a 10 X scope probe. ?
Radio is on 15 m producing approximately 15 watts steady CW output. The scale is 10 V per division, which I verified by briefly measuring the +13 V input. ? 0 V DC is the middle line of the screen. ?
As you can see the voltage across C201 is periodically moving from about +12 to about +20 and back again. It should be solid at 20 V or more I think this is a problem. ? If this is correct then the unit is not keeping the switching diodes back based ? ?I don’t know if my scope probe at 10x is loading it down or what. ? I’m out of time and will have to get back to this. ?
toggle quoted message
Show quoted text
On Nov 10, 2023, at 06:29, Gordon Gibby via groups.io <docvacuumtubes@...> wrote:
? Here I have disconnected the connection to the output diodes but left the loop of wire within a millimeter or two of the solder point so it could pick up capacitively
The output is still pretty clean minus the connection to the output diodes
So I would suggest it has nothing to do with the length or positioning of the wiring. ?
Gordon Kx4z?
On Nov 9, 2023, at 22:18, Gordon Gibby via groups.io <docvacuumtubes@...> wrote:
? Evan, I get what you are going for— that it might be coupling between the loops, however take a look at this post
Where I used to very short jumper across first the input and then the output diodes individually I have the highest low pass filter, with all of the other still connected which is very similar to what you were asking for,…., and it made single-digit changes with still very high harmonics.
I think that suggests that it really does have something to do with all the other systems being connected, possibly their diodes
An easier change that I can do is to simply remove the tack solder point on the output diode (leaving all of the loops of wire still physically present ) and check again to see what spectrum look like. I bet even with the wires still in place but ***no longer soldered ?to the output diodes are the other filters***, it will be again far cleaner
I’ll try to get that done early tomorrow.
Thanks for your review! ?We are closing in on it
73? Gordon
On Nov 9, 2023, at 21:52, Evan Hand <elhandjr@...> wrote:
?Gordon,
Can you remove the long jumper runs for this last test and just jump across D15 and D19?? This would be to eliminate any bypassing of the filters through the jumpers.
73
Evan
AC9TU
|
Hi Gordon,
I agree that your harmonics measurement eliminates the wire as an issue.
The voltage for a 15-watt output should be higher than 40 volts, closer to 70 volts.? The 10x scope (10M Ohms) probe should not load down the doubler.? I think you may have found the issue.? I suggest pulling the diodes to verify that they are OK.? With the diodes removed, you can verify that there are no shorts on the HV line.
If you have a greater than 40-volt bench supply, you could connect it to the HV line to test.? The current requirement would be much less than 1 ma.? If your highest bench supply is 30 volts, reduce the power to less than 8 watts and use it.
By my back-of-the-envelope calculation, the doubler should produce around 52 volts with 15 watts.? You need at least 39 volts (the peak voltage of the transmission plus a diode drop).? I used 40 volts as a buffer and 8 watts for the 30 volts supply.
Check my calculations, and please let me know if I am in error.
73
Evan
AC9TU
|
I haven’t, I haven’t carefully validated your measurements yet. ?
I did change the coupling capacitor from 47 pF to 330 pf— thinking the DC supply isn’t getting enough supply
We are definitely onto something. I’m not sure this “doubler” is doubling
I got my scope to tell me the peak voltage
-Peak AC voltage of RF at the moment is running 25 volts give or take
-peak back bias voltage going to the inductors in the other stages is on the order of 23 to 24.2.?
Obviously, this is not enough backbias!!!!!!!!
Something is bad wrong with the little DC supply in this radio. ?
Gotta go again, but I’ll get back to Gus. I think we are onto something ?thing important ??
Perhaps a diode is backwards or blown up?
toggle quoted message
Show quoted text
On Nov 10, 2023, at 11:19, Evan Hand <elhandjr@...> wrote:
?Hi Gordon,
I agree that your harmonics measurement eliminates the wire as an issue.
The voltage for a 15-watt output should be higher than 40 volts, closer to 70 volts.? The 10x scope (10M Ohms) probe should not load down the doubler.? I think you may have found the issue.? I suggest pulling the diodes to verify that they are OK.? With the diodes removed, you can verify that there are no shorts on the HV line.
If you have a greater than 40-volt bench supply, you could connect it to the HV line to test.? The current requirement would be much less than 1 ma.? If your highest bench supply is 30 volts, reduce the power to less than 8 watts and use it.
By my back-of-the-envelope calculation, the doubler should produce around 52 volts with 15 watts.? You need at least 39 volts (the peak voltage of the transmission plus a diode drop).? I used 40 volts as a buffer and 8 watts for the 30 volts supply.
Check my calculations, and please let me know if I am in error.
73
Evan
AC9TU
|
Gordon,
A Vpeak of 25 volts on a 50 ohm load is P=E^2/100 or 6.25 watts.??
When you get a chance, can you verify with a 50-ohm dummy load the power out of the sBitx?? Something does not compute.
Are you measuring RMS?? That would be P=Vrms^2/50, which is closer to the 15-watt level.? Since we are concerned with the diode state, I would measure the Vpeak, not RMS.? Both are available on most new digital O-scopes.
73
Evan
AC9TU
|
Evan -- the 330 pf capacitor seriously decreased the output power, probably causing what you're seeing. The diodes in the voltage doubler ALL checked out with an ohmmeter (in circuit -- because of the way they are wired, this is possible) as GOOD.? ? Diodes D20 and D21 both checked out as SHORTED.? ? So the RF current through the 47 pf capacitor was being shunted to the +12VDC line (which of course is heavily bypassed to GROUND). I replaced diodes D20 and D21 with (leaded, not SMD) 1N4148's that I had around and bingo!? I now have +60 VDC being developed by that circuit and steady as well.? ?First glance at the spectrum analyzer? indicated a FAR cleaner signal.? ?I have not made photos or extensive measurements, but there is a good chance that THIS WAS THE PROBLEM. I don't know why this would be the problem.? ?This unit has been used a zillion different test scenarios and was purchased second hand also.? ?So I just can't say.? ? None of the diodes was proven to be inserted incorrectly.? ?So I don't know what shorted out D20 and D21.? ? It could have happened at any point.....but I will point out that I started? measuring as soon as I received the radio, before I had done ANY adjustment of the insides.? ?And the previous owner indicated he had never transmitted with it....? ?so I don't? know what was going on.  To help other users, when I get back to it, I will measure with a cheap Harbor Freight VOM and make sure that I can measure that DC voltage there, with RF flowing.? ?? Might there be a way to make more protection of D20/D21?? ?Your comments welcomed, Gordon KX4Z
|
(Note, I also went back to the 47pf capacitor instead of the big 330pf I had put in temporarily).? ??
Wow, this has taken a lot of time to figure out!!!
Gordon KX4Z
|
I'm thinking it might be smart to put some resistance in SERIES with D20/D21.? ?I'm thinking they are there just to keep the LPF switching diodes from being heavily forward biased??? ? Maybe the RF-derived voltage doubler can drive current throughthem?? ?Do diodes fail SHORTED more from over-reverse voltage (avalanche)or from excessive forward current??? ? Maybe 470 ohms in series with them to limit 14volt-based current to ground to 30mA??? ?They are rated to as much as 150mA.? ??
Comments?? ?What causes diodes to? short?
Gordon KX4Z
On Fri, Nov 10, 2023 at 1:38?PM Gordon Gibby via <docvacuumtubes= [email protected]> wrote: (Note, I also went back to the 47pf capacitor instead of the big 330pf I had put in temporarily).? ??
Wow, this has taken a lot of time to figure out!!!
Gordon KX4Z
|
Gordon,
Thinking about your data, I came up with possible reasons for the failure and what the failure is.
Based on the scope measurement of the voltage on C201, I believe one or both D24 and D25 are open.? I base this on your measurement of around 20 volts on C201.? The values would be consistent with a rectified value of the peak voltage.
The failure would be an overvoltage on one, a short on one, then an overvoltage on the other.
73
Evan
AC9TU
|
