Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
Paralleling ProgRock2's
|
A bit of a lesson learned for anyone else who's interested.? I've been working on integrating a ProgRock2 solution into a Heathkit SB-313, to replace a defunct heterodyne oscillator.? The HFO in the SB-313 needs to generate nine frequencies, which is one more than the ProgRock2 can deliver.? I attempted to parallel two ProgRock2's to the same output (with only one powered up at the same time) and if you do that, it blows up the Si5351a chip.? It appears that those clock outputs don't want to see an external signal.? Even if you completely power down the ProgRock2 that's not in use, an external signal on the clock input will terminate the Si5351a.? It's also possible that
As a workaround, I'm considering a SPDT relay on the clock line, but am open to any other thoughts. 73 - Steve, KW4H Tucson, AZ |
|
Steven, It’s always a bad idea to wire-or chip outputs together unless their inputs are being driven by exactly the same signal. Even then, you’re playing with fire tying the outputs together. To accomplish want you want to do, you need some logic after each 5351, essentially nine clock buffers to create 9 individual buffer outputs. Tony AC9QY On Thu, Jul 4, 2024 at 12:41?PM Steven Reed via <reedsteve=[email protected]> wrote: A bit of a lesson learned for anyone else who's interested.? I've been working on integrating a ProgRock2 solution into a Heathkit SB-313, to replace a defunct heterodyne oscillator.? The HFO in the SB-313 needs to generate nine frequencies, which is one more than the ProgRock2 can deliver.? I attempted to parallel two ProgRock2's to the same output (with only one powered up at the same time) and if you do that, it blows up the Si5351a chip.? It appears that those clock outputs don't want to see an external signal.? Even if you completely power down the ProgRock2 that's not in use, an external signal on the clock input will terminate the Si5351a.? It's also possible that |
|
Forgot to mention to never drive active signals into the inputs of a powered-down CMOS chip. Really bad news. On Thu, Jul 4, 2024 at 12:58?PM Tony Scaminaci via <tonyscam=[email protected]> wrote:
|
|
Tony, appreciate the feedback.? In the case of the ProgRock2, it works marvelously well as-is (no external buffers needed) -- however, needing more than 8 frequencies out of a clock output creates these challenges.? Yes, there are two other clock outputs, each with 8 possible frequencies each, but with a single-ended connection (the input to a mixer), a clock output would still see another clock output -- not to mention that at some point there would have to be two clocks running at the same time in a single ProgRock2, which is something I'd rather avoid.? It seems that a SPDT relay that switches the clock output from each ProgRock2 module might do the trick.? What do you think?
73 - Steve, KW4H |
|
Have you tried diode ORing them together?? The diodes should prevent current being pushed back into the chip output.
Or maybe a resistive or transformer-type combiner would work.? This core-and-wire unit has 33dB isolation up to 30MHz, 25 up to 200MHz. ?https://www.minicircuits.com/pdfs/ADP-2-1+.pdf This resistive type works down to DC, with poorer isolation:? It is good practice to avoid putting signals into un-powered ICs, or into outputs. 73, Don N2VGU |
|
Thanks for the tips!? I'm an old school electronics tech -- mostly work on boat anchors, so my work with digital circuits and chips has been fairly limited over the decades. 73 - Steve, KW4H
On Thursday, July 4th, 2024 at 11:33 AM, Donald S Brant Jr <dsbrantjr@...> wrote: Have you tried diode ORing them together?? The diodes should prevent current being pushed back into the chip output. |
|
Hi Steven,
just some thoughts about replacing the HFO. I suppose the SB-313 will be happy with the ~2Vpp square wave on 50 Ohm instead of the original HFO XTAL oscillator output. I would try with simply connecting the outputs of two ProgRock2s using serial 100 Ohm resistors to build a common (HFO) output point. Then I would let both ProgRock2s switched on and selecting the banks in the way that only one frequency would be output. I assume the three MS5351 CLK outputs are switched off if the bank says 0, 0, 0. The other way would be to somehow use the SB-313 HFO mechanical band switch to choose the frequency. -- 73 Bojan S53DZ |
|
Yes -- I know it will work because I've done similar projects for other Heathkit gear.? For clarity, I've already built the bank selection circuitry and modified the band selection wafer switch so it will properly signal the ProgRock2's for the correct frequency for each band.? The switch circuitry also turns on/off each ProgRock2 as needed, so that only one is active.? And Bank 0 is not used.? The first ProgRock2 provides the oscillator signals for the first four bands, and the second ProgRock2 provides the oscillator signals for the remaining five bands.? I have two more ProgRock2's in stock -- so, as you pointed out, would putting a 100 ohm resistor in series with the output of Clock 0 on each ProgRock2 solve the problem?? They have to connect to the same point. 73 - Steve, KW4H
On Thursday, July 4th, 2024 at 1:02 PM, Bojan Naglic <bojan.naglic@...> wrote: Hi Steven, |
|
1. I would think that having a serial resistor (for the first test 200 Ohm?) in each CLK output would not impose a stress to the un-active ProgRock output.
2. Saying that I would certainly not recommend to switch off the power of the unused ProgRock, as Don said before. Just set its CLK output off. 3. I have not tried any of that with the ProgRock. -- 73 Bojan S53DZ |
|
Steve, Relays would work. I completely forgot about your single output requirement…too much 4th partying?. Another possible solution could be an analog multiplexer chip instead of mechanical relays. I saw the other comments about using series resistors in each output for isolation or diodes to prevent back feed into other outputs. These may work but you might still put the other devices at a slight risk. ?There are three rules for CMOS devices: 1. All outputs should be isolated from all other outputs. No feedback allowed. 2. All unused inputs should be tied either to ground or Vdd. Never allow unused inputs to float. 3. Never feed any voltage (other than ground) into an unpowered device. This means no voltage at any inputs when the device isn’t powered. Any method you use to combine multiple outputs into one output should follow these rules. If you use a mechanical relay, ensure that the relay contacts are break before make. Tony? On Thu, Jul 4, 2024 at 1:14?PM Steven Reed via <reedsteve=[email protected]> wrote: Tony, appreciate the feedback.? In the case of the ProgRock2, it works marvelously well as-is (no external buffers needed) -- however, needing more than 8 frequencies out of a clock output creates these challenges.? Yes, there are two other clock outputs, each with 8 possible frequencies each, but with a single-ended connection (the input to a mixer), a clock output would still see another clock output -- not to mention that at some point there would have to be two clocks running at the same time in a single ProgRock2, which is something I'd rather avoid.? It seems that a SPDT relay that switches the clock output from each ProgRock2 module might do the trick.? What do you think? |
|
On 04/07/2024 20:51, Bojan Naglic via groups.io wrote:
having a serial resistor (for the first test 200 Ohm?) in each CLK outputAfter hearing of the damage done to 5351s by connecting the outputs it seems sensible to use buffers. I've not checked but surely there is a way to use one chip which will take the active input from one of two gates? 73 Alan G4ZFQ |
|
In completely digital logic I am using the NOR gates to do a simple 2 input to 1 output mux. I am using it for switching 10MHz signals.
The MS5351 is a digital/analog device with 50 Ohm outputs and the situation is different. But, if the square-wave is not a problem for the SB-313, than also such simple mux would not be. -- 73 Bojan S53DZ |
|
On 05/07/2024 07:05, Bojan Naglic via groups.io wrote:
But, if the square-wave is not a problem for the SB-313Ah yes, I was forgetting we are not feeding a digital receiver. I do not know if the SB-313 will like square waves. Harmonics might not matter, lots of filtering in these old sets. But maybe the relay is the better answer. I'd try to make an analogue switch but that would require a lot of experiments. 73 Alan G4ZFQ |
开云体育A couple of options spring to mind,
Combine the outputs via small diodes like BAT86. A second option is to combine two multiplexer chips
Like 74cbt3253 as a 1 to four mux with a 1 to 4 mux on 3 of the outputs of the first mux. Thus allowing you to route any one of 13 clocks to the input of the first 1 to 4 mux.
There are many ways to sort this, I'd probably use a small CPLD with some routing logic. But the two options above particularly the mux option will work, and are off the shelf parts that need little knowledge to put together.?
Good luck,
Robert G0MYX
Sent from From: [email protected] <[email protected]> on behalf of Alan G4ZFQ <alan4alan@...>
Sent: Friday, July 5, 2024 9:19:59 AM To: [email protected] <[email protected]> Subject: Re: [QRPLabs] Paralleling ProgRock2's ?
On 05/07/2024 07:05, Bojan Naglic via groups.io wrote:
> But, if the square-wave is not a problem for the SB-313 Ah yes, I was forgetting we are not feeding a digital receiver. I do not know if the SB-313 will like square waves. Harmonics might not matter, lots of filtering in these old sets. But maybe the relay is the better answer. I'd try to make an analogue switch but that would require a lot of experiments. 73 Alan G4ZFQ |
开云体育As already suggested you can? get 3 frequencies at a time by using the memory banks on the Progrock 2.
This can be combined with a 4 to 1 Mux 74cbt3253? to give one output selected by the BANK select bits and the the Mux select bits. Only one chip needed with this solution.
Regards,
Robert?
Sent from From: Robert Stephen <rob-stephen@...>
Sent: Friday, July 5, 2024 10:11:29 AM To: [email protected] <[email protected]> Subject: Re: [QRPLabs] Paralleling ProgRock2's ?
A couple of options spring to mind,
Combine the outputs via small diodes like BAT86. A second option is to combine two multiplexer chips
Like 74cbt3253 as a 1 to four mux with a 1 to 4 mux on 3 of the outputs of the first mux. Thus allowing you to route any one of 13 clocks to the input of the first 1 to 4 mux.
There are many ways to sort this, I'd probably use a small CPLD with some routing logic. But the two options above particularly the mux option will work, and are off the shelf parts that need little knowledge to put together.?
Good luck,
Robert G0MYX
Sent from
From: [email protected] <[email protected]> on behalf of Alan G4ZFQ <alan4alan@...>
Sent: Friday, July 5, 2024 9:19:59 AM To: [email protected] <[email protected]> Subject: Re: [QRPLabs] Paralleling ProgRock2's ?
On 05/07/2024 07:05, Bojan Naglic via groups.io wrote:
> But, if the square-wave is not a problem for the SB-313 Ah yes, I was forgetting we are not feeding a digital receiver. I do not know if the SB-313 will like square waves. Harmonics might not matter, lots of filtering in these old sets. But maybe the relay is the better answer. I'd try to make an analogue switch but that would require a lot of experiments. 73 Alan G4ZFQ |
|
After sleeping on this and reading the many fine thoughts and suggestions here, I've decided to try the simplest route:? electrical isolation of the clock output (I only need Clock 0) by way of a SPDT relay.? This also means that it will be unnecessary to power only one ProgRock2 up at a time -- both can be left running.? The other reason for this is that my background is almost entirely in analog electronics, so the use of buffers and multiplexers is best left alone for now.??
73 - Steve, KW4H |
|
I'm going to be using these in an old GE MASTR II repeater I'm refurbishing for 1.25m. In that case, the square wave would be no problem due to the tuning of the multiplier stages following.
I'm also interested in trying one in an Atlas 350XL, since it conveniently has unused crystal sockets. The idea of using sets of 0.5 MHz band segments that are programmable for that radio is enticing because the receiver is very good. 73 Chuck AI7SY |
Donald S Brant Jr
Chuck AI7SY
to navigate to use esc to dismiss