开云体育

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Dave Peterson via groups.io
Sent: Thursday, December 2, 2021 12:17 PM
To: [email protected]
Subject: [TekScopes] TG501 2ns & 1ns operation

Hi all,

I've been working on rehabbing a TG501. After replacing the shorted Tantalum filter caps on the 15v supply the device has operated pretty well. I've been learning the operation, circuits, and tuning of it without too much difficulty. I've gotten the timing marks synced with a Leo Bodnar GPSDO and, despite not having the OPT1 1MHz oscillator, can achieve a very stable output. That is, almost no drift relative to the GPSDO at 1MHz over long periods of time. Nice output levels too.

A remaining issue I'm having leads me to wonder how the 1ns and 2ns circuits are intended to work. First, I understand 1ns is pretty fast for the equipment I have on hand, but I do have a 7934 with 7B92A time base, and 7A29 vertical amplifiers. It's all in less than ideal condition, but functioning well enough for me to observe the 1ns sine wave. Albeit at a weak level. Is this the 7934 BW limit, or the TG501? Undetermined yet.

But the 2ns output is plenty strong - up to 1.5v p2p. Adjustment of C515 between 1ns and 2ns levels has me turning it down to about 800mv p2p. But still I cannot get anything more than about 40mv p2p on the 1ns output (see above - is it the BW limit of the 7934, or the TG501?).

In working on these outputs I've had them both on the 7934 through both the left and right vertical slots, both going through 7A29s. Switching between the 1ns and 2ns push buttons results in both outputs present (2ns on MARKER OUT, 1ns on 1NS ONLY), but when in 1ns mode the 2ns output drops from 800mv p2p (or whatever I've set it to) to about 40mv p2p. The 1ns sine wave output remains fairly consistent at about 40mv p2p. It looks pretty good - a good sine wave, and the level does change a few mV between 1ns and 2ns settings. The 2ns waveform however, goes from looking beautiful in 2ns mode to ugly (not a clean sine waveform) in 1ns mode.

I'm trying to understand how this is happening. Studying the A2 Multiplier Board schematic <4>, I don't see what the 1ns/2ns push buttons even do to affect the 1ns/2ns circuit operation. It looks like these push buttons are tied together at the input of NOR gate U450C. That having either pushed in enables this gate to pass the 100MHz oscillator into the 10ns Amplifier of Q484/Q486. And it also looks like these push buttons both operate/enable K450, and should result in both 2ns and 1ns outputs to be at their respective output BNCs: MARKER OUT and 1NS ONLY. How is selecting 1ns causing the MARKER OUT output to drop?

I'm suspicious of CR515, but given I don't even see how the 10ns Amplifier output should change at all, that seems to be ignorant grasping. Is there something about the input to U450C I'm not understanding? It seems the 1ns and 2ns outputs should always be present, and the 1ns and 2ns push buttons seem redundant - there should just be one button and both outputs?

I'm perfectly comfortable recognizing the 1ns output level to be BW limited by my setup. It makes tuning C515 guesswork, but I can live with that. If I can't even discern the 1ns output reliably with the equipment I have, I don't actually have much (any?) need for it. But it bothers me that that 2ns is so adversely affected. And that I can't understand why the circuit behaves this way. So any insights are greatly appreciated.

Thanks,
Dave

On 12/2/2021 12:17 PM, Dave Peterson via groups.io wrote:

Hi all,

I've been working on rehabbing a TG501. After replacing the shorted Tantalum filter caps on the 15v supply the device has operated pretty well. I've been learning the operation, circuits, and tuning of it without too much difficulty. I've gotten the timing marks synced with a Leo Bodnar GPSDO and, despite not having the OPT1 1MHz oscillator, can achieve a very stable output. That is, almost no drift relative to the GPSDO at 1MHz over long periods of time. Nice output levels too.

A remaining issue I'm having leads me to wonder how the 1ns and 2ns circuits are intended to work. First, I understand 1ns is pretty fast for the equipment I have on hand, but I do have a 7934 with 7B92A time base, and 7A29 vertical amplifiers. It's all in less than ideal condition, but functioning well enough for me to observe the 1ns sine wave. Albeit at a weak level. Is this the 7934 BW limit, or the TG501? Undetermined yet.

But the 2ns output is plenty strong - up to 1.5v p2p. Adjustment of C515 between 1ns and 2ns levels has me turning it down to about 800mv p2p. But still I cannot get anything more than about 40mv p2p on the 1ns output (see above - is it the BW limit of the 7934, or the TG501?).

In working on these outputs I've had them both on the 7934 through both the left and right vertical slots, both going through 7A29s. Switching between the 1ns and 2ns push buttons results in both outputs present (2ns on MARKER OUT, 1ns on 1NS ONLY), but when in 1ns mode the 2ns output drops from 800mv p2p (or whatever I've set it to) to about 40mv p2p. The 1ns sine wave output remains fairly consistent at about 40mv p2p. It looks pretty good - a good sine wave, and the level does change a few mV between 1ns and 2ns settings. The 2ns waveform however, goes from looking beautiful in 2ns mode to ugly (not a clean sine waveform) in 1ns mode.

I'm trying to understand how this is happening. Studying the A2 Multiplier Board schematic <4>, I don't see what the 1ns/2ns push buttons even do to affect the 1ns/2ns circuit operation. It looks like these push buttons are tied together at the input of NOR gate U450C. That having either pushed in enables this gate to pass the 100MHz oscillator into the 10ns Amplifier of Q484/Q486. And it also looks like these push buttons both operate/enable K450, and should result in both 2ns and 1ns outputs to be at their respective output BNCs: MARKER OUT and 1NS ONLY. How is selecting 1ns causing the MARKER OUT output to drop?

I'm suspicious of CR515, but given I don't even see how the 10ns Amplifier output should change at all, that seems to be ignorant grasping. Is there something about the input to U450C I'm not understanding? It seems the 1ns and 2ns outputs should always be present, and the 1ns and 2ns push buttons seem redundant - there should just be one button and both outputs?

I'm perfectly comfortable recognizing the 1ns output level to be BW limited by my setup. It makes tuning C515 guesswork, but I can live with that. If I can't even discern the 1ns output reliably with the equipment I have, I don't actually have much (any?) need for it. But it bothers me that that 2ns is so adversely affected. And that I can't understand why the circuit behaves this way. So any insights are greatly appreciated.

Thanks,
Dave

On Thursday, December 2, 2021, 12:25:40 PM PST, Raymond Domp Frank <hewpatek@...> wrote:

On Thu, Dec? 2, 2021 at 08:52 PM, Dave Peterson wrote:

Why does switching 2ns/1ns push buttons change the 2ns output at all!? Given
the schematic for the 10ns Amplifier input U450C.

S350-1 and S350-2 are closed in the 1-2-5 ns position.
With collector of Q270 low (Out of Lock filter), S450B (2ns) connects Marker Out to C515 (2ns signal) via K450-S1 K450.
With 1 ns button depressed, K450-S1 opens that link, reducing and changing load on 2ns filter output and 2ns filter is coupled only to 1ns filter.

Raymond

On 12/2/2021 5:51 PM, Dave Peterson via groups.io wrote:

I guess my next question would be: is that horrible 40mV p2p signal coming out of the MARKER OUT "normal" in 1ns mode? Yuck.

I can see how mitigating coupling through the contacts of K450 could be challenging. But does this indicate any issues with Q460, CR468, or any other associated components and wiring? It seems that Q460 should be pulling MARKER OUT up.

On Thursday, December 2, 2021, 03:31:39 PM PST, Harvey White <madyn@...> wrote:

IIRC, the marker is labeled 2ns to (whatever), which makes me suspect
that there was never intended to be a marker out at 1 ns.? Haven't
looked at the output, though.

Harvey


On 12/2/2021 5:51 PM, Dave Peterson via groups.io wrote:

I guess my next question would be: is that horrible 40mV p2p signal coming out of the MARKER OUT "normal" in 1ns mode? Yuck.

I can see how mitigating coupling through the contacts of K450 could be challenging. But does this indicate any issues with Q460, CR468, or any other associated components and wiring? It seems that Q460 should be pulling MARKER OUT up.

On 12/2/2021 6:41 PM, Dave Peterson via groups.io wrote:

Nope, not intended nor expected. But I don't expect to see the MARKER OUT output to be putting out anything while in 1ns mode. The 1ns output on the 1ns ONLY BNC is otherwise fine - as good as my equipment can make out.
The MARKER OUT should be pulled up by Q460, and I'm seeing a crappy 40mV ~2ns (500MHz) sine wave. I've taken a look at Q460. It's not socketed, so I'm not going to pull it at this point to check on a curve tracer. Not yet anyway. I would like to know what is expected. Am I right that it should be near +5v (minus a diode drop?)?

On Thursday, December 2, 2021, 03:31:39 PM PST, Harvey White <madyn@...> wrote:
IIRC, the marker is labeled 2ns to (whatever), which makes me suspect
that there was never intended to be a marker out at 1 ns.? Haven't
looked at the output, though.

Harvey


On 12/2/2021 5:51 PM, Dave Peterson via groups.io wrote:

I guess my next question would be: is that horrible 40mV p2p signal coming out of the MARKER OUT "normal" in 1ns mode? Yuck.

I can see how mitigating coupling through the contacts of K450 could be challenging. But does this indicate any issues with Q460, CR468, or any other associated components and wiring? It seems that Q460 should be pulling MARKER OUT up.

On Thursday, December 2, 2021, 04:18:29 PM PST, Raymond Domp Frank <hewpatek@...> wrote:

On Fri, Dec? 3, 2021 at 12:41 AM, Dave Peterson wrote:

Nope, not intended nor expected. But I don't expect to see the MARKER OUT
output to be putting out anything while in 1ns mode. The 1ns output on the 1ns
ONLY BNC is otherwise fine - as good as my equipment can make out.
The MARKER OUT should be pulled up by Q460, and I'm seeing a crappy 40mV ~2ns
(500MHz) sine wave. I've taken a look at Q460. It's not socketed, so I'm not
going to pull it at this point to check on a curve tracer. Not yet anyway. I
would like to know what is expected. Am I right that it should be near +5v
(minus a diode drop?)?

I'm not sure if you realize that 2ns and 1ns aren't generated by the 5ns+ circuits; they are separated by K450-S1, which selects either all 5ns+ signals or both the 1ns and 2ns signals. The fact that you can see 1,2 ns and 5ns+ signals indicates that K450-S1 does its selection job. For 1ns and 2ns, everything "above" K450-S1, i.e. from P468, doesn't count.

The 2ns signal is the bandwidth-filtered result of a dirty signal going into the 2ns filter (CL510, C/L511, C/L512, CL513).
CR515 "sharpens" that signal, creating harmonics, which are then bandwidth-filtered out by C/L517, C/L518, C/L519 and C/L5120. the result is a "soft" signal with a frequency of 1GHz, presented at the "1ns only" BNC socket.

The irregular 2ns signal in your picture looks like the result of mixing, caused by leakage: Bad screening by the aluminium cage covering the 1ns and 2ns etc. circuits may cause this. It should be a pretty tight and complete fit. Just the screws holding it in place and making contact isn't enough.
If you can make a setup where you can put pressure on the cage with power on, I wouldn't be surprised if the signal shape of 1ns and 2ns signals visibly varies, depending on pressure exerted on the cage cover.
If that is so, you may want to remove the cage lid and bend its sides together, even to a slightly smaller than 90 degree angle. Use some tool to fit the lid on and put in the screws.
Next, perform the adjustment for the 2ns and 1ns signals (C/L510 ...C/L520), unless everything looks perfect already. Preferably use a ceramic screwdriver, definitely not one with a metal tip. Simple peaking of the signal and watching for a regular signal on your scope will do the job; first 2ns, then 1ns, and repeat until no more changes. This usually is enough to achieve excellent results. The BW of your 'scope is only important in so far as it is able to show and trigger on the signal. Correct amplitude is not important.

Raymond

On 12/2/2021 17:47, Dave Peterson via groups.io wrote:

The schematic depiction of switches in Tek schematics isn't always clear to me.

Yes, I was clear on the 5, 2, and 1 ns circuits - really the 5ns and 1/2ns circuits, and that K450 switches between them. Thus my confusion regarding K450 DE-activating in 1ns mode. I further missed that in NOT 5ns mode the base of Q460 is pulled up by the disconnection of the 5ns pushbutton (S450A) and R451. And checking the 2ns and 5ns output I now understand why the 5ns mode jumps up in DC offset compared to the 2ns output. Incremental understanding.

So the 2ns Filter output is disconnected from MARKER OUT, Q460 is off, anode of CR468 is pulled down by R465, and MARKER OUT is floating. That explains is susceptibility to noise. Recall I'm the guy who has Sutro Tower looming over my house blasting me with 100MHz broadcast frequencies. Still, the proximity of the 500MHz/2ns signal out of the 2ns Filter is plenty to appear on the MARKER OUT.

I had (in the pictures) the 2ns level C515 set for a large output. Adjusting it to a minimum 2ns level results in a significantly smaller "noise" level on MARKER OUT in 1ns mode. No surprise. There also seems to be a balance between maximum 2ns level and maximum 1ns level. As this is all a tuned resonant type circuit, I'm not surprised. I'm working on rebuilding my small signal analysis chops. And again, I'll say I'm closer to full understanding. But it's always a closer approximation - never quite fully there.

As an aside, since you brought it up: I've been looking for a suitable non-capacitive trimming screwdriver. I just tried making one out of hard plastic, but these aged devices tend to be kind of sticky. It can't stand up to the torque needed to turn some of the caps. So I have to resort to a metal trimmer to get the caps turned. I can do incremental adjustments and get the effects I need.

I've been wary of buying plastic non-capacitive trimmers online. I suspect they'll be no better than I've been able to make myself. You mentioned ceramic. Do you have a particular brand or source for such? An Amazon link perhaps?

Thanks.Dave


On Thursday, December 2, 2021, 04:18:29 PM PST, Raymond Domp Frank <hewpatek@...> wrote:
On Fri, Dec? 3, 2021 at 12:41 AM, Dave Peterson wrote:

Nope, not intended nor expected. But I don't expect to see the MARKER OUT
output to be putting out anything while in 1ns mode. The 1ns output on the 1ns
ONLY BNC is otherwise fine - as good as my equipment can make out.
The MARKER OUT should be pulled up by Q460, and I'm seeing a crappy 40mV ~2ns
(500MHz) sine wave. I've taken a look at Q460. It's not socketed, so I'm not
going to pull it at this point to check on a curve tracer. Not yet anyway. I
would like to know what is expected. Am I right that it should be near +5v
(minus a diode drop?)?

I'm not sure if you realize that 2ns and 1ns aren't generated by the 5ns+ circuits; they are separated by K450-S1, which selects either all 5ns+ signals or both the 1ns and 2ns signals. The fact that you can see 1,2 ns and 5ns+ signals indicates that K450-S1 does its selection job. For 1ns and 2ns, everything "above" K450-S1, i.e. from P468, doesn't count.

The 2ns signal is the bandwidth-filtered result of a dirty signal going into the 2ns filter (CL510, C/L511, C/L512, CL513).
CR515 "sharpens" that signal, creating harmonics, which are then bandwidth-filtered out by C/L517, C/L518, C/L519 and C/L5120. the result is a "soft" signal with a frequency of 1GHz, presented at the "1ns only" BNC socket.

The irregular 2ns signal in your picture looks like the result of mixing, caused by leakage: Bad screening by the aluminium cage covering the 1ns and 2ns etc. circuits may cause this. It should be a pretty tight and complete fit. Just the screws holding it in place and making contact isn't enough.
If you can make a setup where you can put pressure on the cage with power on, I wouldn't be surprised if the signal shape of 1ns and 2ns signals visibly varies, depending on pressure exerted on the cage cover.
If that is so, you may want to remove the cage lid and bend its sides together, even to a slightly smaller than 90 degree angle. Use some tool to fit the lid on and put in the screws.
Next, perform the adjustment for the 2ns and 1ns signals (C/L510 ...C/L520), unless everything looks perfect already. Preferably use a ceramic screwdriver, definitely not one with a metal tip. Simple peaking of the signal and watching for a regular signal on your scope will do the job; first 2ns, then 1ns, and repeat until no more changes. This usually is enough to achieve excellent results. The BW of your 'scope is only important in so far as it is able to show and trigger on the signal. Correct amplitude is not important.

Raymond

-------- Original message --------From: Tom Lee <tomlee@...> Date: 12/2/21 6:02 PM (GMT-08:00) To: [email protected] Subject: Re: [TekScopes] TG501 2ns & 1ns operation Re: non-conductive toolsI often make my own from cast-off bits of FR4. A little filing action, and you've got your tool. Works great.--Tom-- Prof. Thomas H. LeeAllen Ctr., Rm. 205350 Jane Stanford WayStanford UniversityStanford, CA 94305-4070 12/2/2021 17:47, Dave Peterson via groups.io wrote:>?? The schematic depiction of switches in Tek schematics isn't always clear to me.>> Yes, I was clear on the 5, 2, and 1 ns circuits - really the 5ns and 1/2ns circuits, and that K450 switches between them. Thus my confusion regarding K450 DE-activating in 1ns mode. I further missed that in NOT 5ns mode the base of Q460 is pulled up by the disconnection of the 5ns pushbutton (S450A) and R451. And checking the 2ns and 5ns output I now understand why the 5ns mode jumps up in DC offset compared to the 2ns output. Incremental understanding.>> So the 2ns Filter output is disconnected from MARKER OUT, Q460 is off, anode of CR468 is pulled down by R465, and MARKER OUT is floating. That explains is susceptibility to noise. Recall I'm the guy who has Sutro Tower looming over my house blasting me with 100MHz broadcast frequencies. Still, the proximity of the 500MHz/2ns signal out of the 2ns Filter is plenty to appear on the MARKER OUT.>> I had (in the pictures) the 2ns level C515 set for a large output. Adjusting it to a minimum 2ns level results in a significantly smaller "noise" level on MARKER OUT in 1ns mode. No surprise. There also seems to be a balance between maximum 2ns level and maximum 1ns level. As this is all a tuned resonant type circuit, I'm not surprised. I'm working on rebuilding my small signal analysis chops. And again, I'll say I'm closer to full understanding. But it's always a closer approximation - never quite fully there.>> As an aside, since you brought it up: I've been looking for a suitable non-capacitive trimming screwdriver. I just tried making one out of hard plastic, but these aged devices tend to be kind of sticky. It can't stand up to the torque needed to turn some of the caps. So I have to resort to a metal trimmer to get the caps turned. I can do incremental adjustments and get the effects I need.>> I've been wary of buying plastic non-capacitive trimmers online. I suspect they'll be no better than I've been able to make myself. You mentioned ceramic. Do you have a particular brand or source for such? An Amazon link perhaps?>> Thanks.Dave>>>????? On Thursday, December 2, 2021, 04:18:29 PM PST, Raymond Domp Frank <hewpatek@...> wrote:>?? >?? On Fri, Dec? 3, 2021 at 12:41 AM, Dave Peterson wrote:>>> Nope, not intended nor expected. But I don't expect to see the MARKER OUT>> output to be putting out anything while in 1ns mode. The 1ns output on the 1ns>> ONLY BNC is otherwise fine - as good as my equipment can make out.>> The MARKER OUT should be pulled up by Q460, and I'm seeing a crappy 40mV ~2ns>> (500MHz) sine wave. I've taken a look at Q460. It's not socketed, so I'm not>> going to pull it at this point to check on a curve tracer. Not yet anyway. I>> would like to know what is expected. Am I right that it should be near +5v>> (minus a diode drop?)?>>> I'm not sure if you realize that 2ns and 1ns aren't generated by the 5ns+ circuits; they are separated by K450-S1, which selects either all 5ns+ signals or both the 1ns and 2ns signals. The fact that you can see 1,2 ns and 5ns+ signals indicates that K450-S1 does its selection job. For 1ns and 2ns, everything "above" K450-S1, i.e. from P468, doesn't count.>> The 2ns signal is the bandwidth-filtered result of a dirty signal going into the 2ns filter (CL510, C/L511, C/L512, CL513).> CR515 "sharpens" that signal, creating harmonics, which are then bandwidth-filtered out by C/L517, C/L518, C/L519 and C/L5120. the result is a "soft" signal with a frequency of 1GHz, presented at the "1ns only" BNC socket.>> The irregular 2ns signal in your picture looks like the result of mixing, caused by leakage: Bad screening by the aluminium cage covering the 1ns and 2ns etc. circuits may cause this. It should be a pretty tight and complete fit. Just the screws holding it in place and making contact isn't enough.> If you can make a setup where you can put pressure on the cage with power on, I wouldn't be surprised if the signal shape of 1ns and 2ns signals visibly varies, depending on pressure exerted on the cage cover.> If that is so, you may want to remove the cage lid and bend its sides together, even to a slightly smaller than 90 degree angle. Use some tool to fit the lid on and put in the screws.> Next, perform the adjustment for the 2ns and 1ns signals (C/L510 ...C/L520), unless everything looks perfect already. Preferably use a ceramic screwdriver, definitely not one with a metal tip. Simple peaking of the signal and watching for a regular signal on your scope will do the job; first 2ns, then 1ns, and repeat until no more changes. This usually is enough to achieve excellent results. The BW of your 'scope is only important in so far as it is able to show and trigger on the signal. Correct amplitude is not important.>> Raymond>>>>>>??? >>> >>