P6046 attenuator head manufacture

[SArepairman]

http://www.ebay.com/itm/TEKTRONIX-P6046-10X-ATTENUATOR-010-0419-00-/191136808588?pt=LH_DefaultDomain_0&hash=item2c80a43e8c

I have a P6046 but I don't have the attenuator head. I don't want to pay for that. Any idea what goes into making a little attenuator head?


what is the 2puff capacitor connected to, it looks like one side is floating and the signal is routed between the plates of a air capacitor?

I also guess that finding matched 9M resistors might be a challange.

does that mean shielded 9M resistors?

there is no reason I could not use 10M resistors, so long I don't mind a slightly odd attenuation value right? But TBH if I am measuring a 250V signal, I don't really care if its 9:1 or 11:1, its gonna be rough

[Vgkid]

It looks like the 2pf cap is in parallel to the 9 meg resistor. If you changed it to a 10 meg it would throw off the attenuation as you noted. Though you could use 3, 3 meg resistors in series. What frequencies are you looking at?

[SArepairman]

It looks like the 2pf cap is in parallel to the 9 meg resistor. If you changed it to a 10 meg it would throw off the attenuation as you noted. Though you could use 3, 3 meg resistors in series. What frequencies are you looking at?

why the hell did they draw it so weird if its just a parallel cap?

I wanted to get as much frequency as possible. I noticed vishay samples 25PPM 9.1M 1% resistors mismatch would result in greater CMRR error. I wonder what they are matched to in the real attenuator. But realistically at these high voltages I can't imagine it would be more then 1-2MHz.

could those capacitors be connected to the shield?
I mean its hand written too, I don't think its a formatting error.

[Sailor]

They're drawn like a feed-through cap, except that while the normal case is to have one side in contact with the enclosure that you are feeding the signal/power/whatever into or out-of, here they appear to connect it towards the signal output Strange...

[Vgkid]

Looking at the input impedance specs, the accuracy is +- 2%. It would be easier to use a string of 3 meg resistors to get to the desired voltage rating.
These might work 1%, but you could select the ones closest to the 3 meg mark.
http://www.mouser.com/ProductDetail/Vishay-Beyschlag/MBB02070C3004FC100/?qs=sGAEpiMZZMu61qfTUdNhG%252bgK%2fIhJPI67nFj3Hnpdj1c%3d

[SArepairman]

i wonder if the guy who drew that thought in 50 years internet archeologists will be perplexed by me

And yea, you guys are right its much easier to get 3+ resistors then 1 9M resistor. 9M seems like an obscure value.

the 3M are only 20 cents so you can buy a bunch and ratio match them nicely for high  CMRR

[SArepairman]

i wonder if the guy who drew that thought in 50 years internet archeologists will be perplexed by me

And yea, you guys are right its much easier to get 3+ resistors then 1 9M resistor. 9M seems like an obscure value.

the 3M are only 20 cents so you can buy a bunch and ratio match them nicely for high  CMRR

how about the trim pots, any suggestions?

http://www.mouser.com/Passive-Components/Capacitors/Trimmer-Variable-Capacitors/_/N-5g94/

I am not even sure how to determine the qualities of a good trimmer, I know alot about adjustable resistors but nothing of adjustable capacitors.

[Vgkid]

There are more choices for the 3-10pf range, here are the through hole choices
http://www.mouser.com/Passive-Components/Capacitors/Trimmer-Variable-Capacitors/_/N-5g94Zscv7?P=1yzeg60Z1z0wm9sZ1z0z7l5Z1z0z80h&Ns=Pricing%7c1

* I'm leaving out the pricey multiturn caps   

http://www.mouser.com/Passive-Components/Capacitors/Trimmer-Variable-Capacitors/_/N-5g94?Keyword=multi+turn

[SArepairman]

I think I have a little metal toy hot-wheels van that this circuit will fit into nicely 

we just need some grey beard to confirm wtf those capacitors are doing.

I wonder if they did something funny, like run a wire from behind the resistor parallel to the resistor in order to create some kind of distributed capacitance ...
there has to be more to that picture.

[PA0PBZ]

But TBH if I am measuring a 250V signal, I don't really care if its 9:1 or 11:1, its gonna be rough

I think the max voltage is 200mV without the attenuator, so 2V with it in place.

what is the 2puff capacitor connected to, it looks like one side is floating and the signal is routed between the plates of a air capacitor?

Looks like it is just the input capacitance, it has a metal sleeve around the input pin.

[BravoV]

P6046 has 25 Volt "absolute" max input ( + peak AC) and with 5 V common mode range, while with the attenuator only increases ten fold.

Looking at your requirement to probe at 250 Volt, this level is not safe for your probe there.

Edit : Quoting from the P6046 manual

"Its suggested that points to be checked be tested with "ANOTHER" device to insure that voltages do not exceed the P6046 probe capabilites."

CAUTION
Breakdown voltage is +/- 25 V, DC + peak AC, extends to 250 V when the Probe Dual Attenuator Head is installed.

[Richard Head]

It looks like simply two 10:1 scope probes in one package.
This is how it works:
The 9M series resistor and the scope 1M input resistance form a 10:1 potential divider for DC and low frequency AC signals.
However, the scope input capacitance starts causing slew rate limiting due to the RC time constant of the 9M resistor and the 20pF input capacitance of the vertical amplifier.
To solve the problem a parallel capacitor is placed across the series 9M resistor and an additional capacitor across the scope input. Note that the ratio of the total parallel capacitance (Cinput +Ctrim) to series capacitance must also be 9:1 to have the same division ratio as the resistors. The capacitor across the scope input is made adjustable so that the exact division ratio can be established. This is the cap you adjust to produce a nice square sided waveform on the scope screen.

Dick

[Richard Head]

It looks like simply two 10:1 scope probes in one package.
This is how it works:
The 9M series resistor and the scope 1M input resistance form a 10:1 potential divider for DC and low frequency AC signals.
However, the scope input capacitance starts causing slew rate limiting AS THE FREQUENCY INCREASES due to the RC time constant of the 9M resistor and the 20pF input capacitance of the vertical amplifier.
To solve the problem a parallel capacitor is placed across the series 9M resistor and an additional capacitor across the scope input. Note that the ratio of the total parallel capacitance (Cinput +Ctrim) to series capacitance must also be 9:1 to have the same division ratio as the resistors. The capacitor across the scope input is made adjustable so that the exact division ratio can be established. This is the cap you adjust to produce a nice square sided waveform on the scope screen.

Dick

[David Hess]

The P6046 lacks input protection to keep its input capacitance low and AC common mode rejection high.

The adjustable capacitors in the attenuator head are to both compensate the voltage divider and trim the AC common mode rejection when the attenuator is attached.

I think the feedthrough capacitors shown are more like controlled parasitic capacitance or shielded resistors which could be also shown using a line parallel to the resistor and connected to one side.  The input capacitance on each side needs to match for high AC common mode rejection.  Swamping the natural parasitic capacitance of each resistor allows it to be tightly controlled.  Shielding them with the shield connected on the probe side has this effect. 

[SArepairman]

Hmm, so how should I construct this circuit?  I was thinking to use all through hole parts and do a 2-d dead-bug lay out.

I wanted to put it in my hot-wheels van and drill holes that would allow me to insert a ceramic adjustment rod for the trim capacitors. I don't think I will be able to get my hands on the nice tektronix oscope probe style sharp tip, I was just going to solder gold pins to the 9M resistors, the type used in the center connector for a larger type coaxial cable "aimed" at each other so they can easily probe across a SMD resistor, specifically for current measurements.



So, am I doing anything about that parallel 2puff capacitance? I am getting too many conflicting replies in this thread. And where would I connect the van to the ground point? Should I use a single wire? many wires?

[David Hess]

Just to keep things simple, I would construct the circuit without the shielded resistors and make it as physically symmetrical as possible.

The next step up from that which I would try, beside using surface mount construction to minimize parasitics, is to make an adjustable wire gimmick (look it up) to duplicate the shielding on the resistors.  Trimming the AC common mode rejection is not a trivial task however.

[SArepairman]

Just to keep things simple, I would construct the circuit without the shielded resistors and make it as physically symmetrical as possible.

The next step up from that which I would try, beside using surface mount construction to minimize parasitics, is to make an adjustable wire gimmick (look it up) to duplicate the shielding on the resistors.  Trimming the AC common mode rejection is not a trivial task however.

ah I looked up the wire gimmick , its exactly what I suspected from the schematic, parallel wire.

you say its tricky because I would need to adjust the "turns ratio" of the gimmick and the ratio of the trim caps, so its like a equation with two variables? I could probably get good matches on resistors. 

[David Hess]

Just to keep things simple, I would construct the circuit without the shielded resistors and make it as physically symmetrical as possible.

The next step up from that which I would try, beside using surface mount construction to minimize parasitics, is to make an adjustable wire gimmick (look it up) to duplicate the shielding on the resistors.  Trimming the AC common mode rejection is not a trivial task however.

ah I looked up the wire gimmick , its exactly what I suspected from the schematic, parallel wire.

In this case I would just run the wire parallel to the resistor instead of using a second wire and then trim or bend the wire to make the adjustment.

you say its tricky because I would need to adjust the "turns ratio" of the gimmick and the ratio of the trim caps, so its like a equation with two variables? I could probably get good matches on resistors.

The AC adjustment of the P6046 probe and attenuator is known to be cumbersome.  Special test fixtures are practically required.  That is one of the reasons I would initially keep it simple as the calibration required to take advantage of a better design is difficult at best.