|
Hi all,?
I’m considering using my TinySA Ultra as a signal generator to fault find a VHF SSB transceiver, I want a stable 10.7MHz signal that I can inject into the signal chain but I’m worried about signal levels and damaging the TinySA output, I will obviously be using a blocking capacitor.
can anyone give me a heads up on how to do this please?
regards
Amanda M0DZO
|
On Wed, Feb 7, 2024 at 06:48 AM, Amanda Hambidge wrote:
can anyone give me a heads up on how to do this please?
Personally, I'd place an extra external attenuator on the generator output, and then use a DC blocking capacitor. If you are not worried about absolute signal levels, then put something like a 1k ohm resistor in series with the blocking capacitor, as this will help to further protect the attenuator and signal generator. It will also help to reduce the loading effects of the 50 ohm source, whilst you are poking around the circuit trying to trace the signal path. Another method, if you have one handy, is to use an X10 scope probe to inject the signal. Good luck. Regards, Martin
|
Hi Martin,?
thanks for the reply, I never thought of the x10 scope Probe idea, I do have signal generators but it’s the time waiting for them to stabilise that can be a pain.
regards
Amanda
|
On Wed, Feb 7, 2024 at 12:49 PM, Amanda Hambidge wrote:
I never thought of the x10 scope Probe idea
You will get about 20 to 30dB loss through a X10 probe, as they are designed for use with a high input impedance scope. But if you are not bothered about the accuracy of the signal amplitude, it's a very quick and handy way of injecting a low level signal. Regards, Martin
|
+1 on the 10X scope probe. ?
|
Actually the figure is about -52.5dB. 9 MegOhm/50 Ohm divider.
10log(1.8 x 10minus5).
73...Bob VK2ZRE
On 8/02/2024 2:29 am, Martin via
groups.io wrote:
toggle quoted message
Show quoted text
On Wed, Feb 7, 2024 at 12:49 PM, Amanda Hambidge wrote:
I never thought of the x10 scope Probe idea
You will get about 20 to 30dB loss through a X10 probe, as they
are designed for use with a high input impedance scope.
But if you are not bothered about the accuracy of the signal
amplitude, it's a very quick and handy way of injecting a low
level signal.
Regards,
Martin
|
On Thu, Feb 8, 2024 at 12:40 PM, Bob Ecclestone wrote:
Actually the figure is about -52.5dB. 9 MegOhm/50 Ohm divider.
10log(1.8 x 10minus5).
Hi Bob, Yes, I'd agree on these theoretical values, but a lot depends upon the probe construction and the load impedance, and I have successfully used this technique in practice for many years. Don't forget that although they may use a 9M ohm resistor as the voltage divider, it also has a frequency compensation capacitor in parallel with it, and this is usually around 10pF.  The figures I originally quoted were ones I'd quickly measured into a 50 ohm load impedance. I have just tried three different makes, and they all show different levels of attenuation when used in the way I suggested. The tracking generator output was set to 0dBm and the frequency scale is from 0-1.5GHz.  I believe this to be good enough for the suggested purpose. Regards, Martin
|
One other thing I forgot to mention.
You can use a X10 scope probe with a spectrum analyser and will obtain a similar level of attenuation.
This can be handy, for example, when tracing an RF signal through a series of amplifier stages. But be careful that the RF level isn't excessive if you are dealing with high power devices.
If in doubt, put an external attenuator on the SA input, just to be on the safe side.
Regards,
Martin
|
And another thing...
If you want to build your own passive probe, then using a series connected resistor and capacitor is a better option.
The principle is described on this webpage.
950 Ohms (330 & 620 ohms in series) and a 0.1uF capacitor with a suitable voltage rating, will provide about 20dB attenuation, assuming a 50 ohm source and load.
Placing a couple of 1N4148 diodes connected in antiparallel across the probe output where it connects to the SA input is another useful safety precaution.
Regards,
Martin
Regards,
Martin
|
Hi Martin et al,
thanks very much for the information it’s very much appreciated, as I said my main worry was blowing up the output of the rf generator, the scope probe is definitely one I will try. Lots of useful info, thanks again.
Amanda M0DZO
|
I would not put 1n4148 diodes in there, 2 of them anti parallel into 50 ohm is roughly +10dbm. While that not always destroy the TinySA it will be past safe limits. I would say 2 low capacity? diodes like the BAT45 or similar would be the better option. With it, the maximum power into 50 ohm is only about +5dbm
|
Hi Martin,
You are absolutely correct. I thought about it a bit more pretty
much as soon as I pressed "send'.
I have now extracted my foot from my mouth and have put my teeth
back in:-)
Firstly, there are two modes:
1. SA Mode where probe is used on the input to the TinySA. I believe
the -52dB is pretty close in this case. But I will check.
Again, you are correct that the compensation capacitor will make the
probe attenuation frequency dependant, depending on the input
capacitance of the TinySA.
2. Sig Gen Mode where the probe is used to inject a signal into a
circuit.
Here the impedance at the injection point will most likely be much
higher than 50 ohms and therefore the attenuation will be less,
probably much closer to your 20-30dB.
I can't argue with you practical experience here. I have also used a
probe in similar circumstances but never measured the results.
I recently purchased a handheld digital oscilloscope and had
intended doing some frequency response measurements on it using a
TinySA as a generator and the TinySA Ultra to measure the level with
the CRO bridged across the Ultra input.
I had also intended to check a few discrete frequencies using the
x10 probe on the input to the Ultra. 455KHz, 10.7MHz for peaking IFs
and 147MHz. The CRO is supposed to be good to 180MHz. I have already
checked that I can see 147MHz from the TinySA in generator mode and
I can. I did not have an SMA "T" at the time, so I could not do the
3 way test.
The use of the oscilloscope probe is a very convenient means of
connecting to a circuit under investigation, both in probing and
injecting a signal.
Still a work in progress.
Cheers...Bob VK2ZRE
On 9/02/2024 1:14 am, Martin via
groups.io wrote:
toggle quoted message
Show quoted text
On Thu, Feb 8, 2024 at 12:40 PM, Bob Ecclestone wrote:
Actually the figure is about -52.5dB. 9 MegOhm/50 Ohm
divider.
10log(1.8 x 10minus5).
Hi Bob,
Yes, I'd agree on these theoretical values, but a lot depends upon
the probe construction and the load impedance, and I have
successfully used this technique in practice for many years.
Don't forget that although they may use a 9M ohm resistor as the
voltage divider, it also has a frequency compensation capacitor in
parallel with it, and this is usually around 10pF.
The figures I originally quoted were ones I'd quickly measured
into a 50 ohm load impedance.
I have just tried three different makes, and they all show
different levels of attenuation when used in the way I suggested.
The tracking generator output was set to 0dBm and the frequency
scale is from 0-1.5GHz.
I believe this to be good enough for the suggested purpose.
Regards,
Martin
|
And don't forget the DC Block, particularly if working on tube
(valve) equipment.
Cheers...Bob VK2ZRE
On 9/02/2024 1:39 am, Martin via
groups.io wrote:
toggle quoted message
Show quoted text
One other thing I forgot to mention.
You can use a X10 scope probe with a spectrum analyser and will
obtain a similar level of attenuation.
This can be handy, for example, when tracing an RF signal through
a series of amplifier stages. But be careful that the RF level
isn't excessive if you are dealing with high power devices.
If in doubt, put an external attenuator on the SA input, just to
be on the safe side.
Regards,
Martin
|
There has been quite a thread on this forum on the construction of a probe that limits the signal level into the spectrum analyser, I built one based on the design by Marek, notes and details of both are in the files section of this forum. It is powered from USB, or any 5V supply. 73 Dave 
|
On Fri, Feb 9, 2024 at 08:21 AM, G4AON wrote:
the design by Marek
Marek's design is very good if you actually require an active probe. They have their uses, but in my experience in most cases, you don't require the additional sensitivity or high input impedance that they can provide. A simple passive probe is adequate, plus you can inject signals with it too. However, it is worthwhile having an active probe on hand, as in the very few occasions you may need one, they are invaluable. Like everything else, the more tools you have, the easier the jobs become. Regards, Martin
|
Hi Dave,?
not being in the know about pcb files, how do I go about sending the info to the PCB manufacturer you used?
regards
Amanda M0DZO
|
On Fri, Feb 9, 2024 at 04:46 AM, Bob Ecclestone wrote:
I have now extracted my foot from my mouth and have put my teeth back in:-)
Hi Bob, Ha, you and me too. The number of times I've pressed send and then immediately had a Doh moment... Someone once said "you learn more from your mistakes" if that were true I'd be competing with Einstein. Regards, Martin
|
On Fri, Feb 9, 2024 at 01:16 AM, John Cunliffe W7ZQ wrote:
I would say 2 low capacity? diodes like the BAT45 or similar would be the better option
Hi John, Yes I'd agree. My thinking was to use parts that folks were likely to have lying around, and 1N4148's seemed to be a better option than nothing. It was to try and limit short duration spikes due to the coupling capacitor charging / discharging, more so than RF limiting, but I take your point. Regards, Martin
|
On Fri, Feb 9, 2024 at 11:43 AM, Amanda Hambidge wrote:
Hi Dave,?
not being in the know about pcb files, how do I go about sending the info to the PCB manufacturer you used?
regards
Amanda M0DZO
I am away from home until Monday, but there is a chance I have a spare board I can post you FOC. I will drop you an email via your QRZ email. 73 Dave
|
Martin, I agree, the 1N4148 is the diode that is probably in every junk box and will work. I just thought about trouble shooting I have done in the past especially on tube receivers and TV audio IF in the late 60s and 70s (yes I am dating myself here lol) It is just too easy to touch the wrong pin and suddenly have the full local oscillator signal going into the probe.For sure strong enough to generate a 10dbm or more signal at the input of the TinySA if using the 4148s.I have used the BAT45 many a times in similar applications,it's low capacity makes it a good candidate for higher frequencies. However, they are hard to get these days unless one buys some questionable ones from E-bay. But there are ample of similar low capacity small signal schottky diodes available for people to chose from.
John
|
John Cunliffe W7ZQ
G4AON