CurtPalme.com Home Theater Forum

[Nashou66]

[dvh99]

thanks nash i will look at that later, now i am off playing some soccer .

[tse]

http://www.fairchildsemi.com/ds/MP/MPSH81.pdf

might work ok.

I used the 2N3906 on that proto board and had good results.

Scott

[dvh99]

http://nl.farnell.com/nxp/bfr93/transistor-npn-rf-sot-23/dp/1081298?Ntt=bfr93

you mean this one nash? i want those they were manufactured here at philips .

scott i am looking for something with a higher bandwidth then 250mhz.
what was the idea behind developing a vim with the ad834, the ad835 was good enough for most installations and still is.

[Nashou66]

[tse]

[1031]

[dvh99]

jarmo the clc449 is not available anymore, i will order the el5166 too but i like the ths3201 to work.
farnell also has special rf caps that i will try on the ths3201 to see how that works maybe 5 in parallel.
how does the 2n3906 perform with pixelclocks above 300 mhz?
about the convergence scott, it isnt the best but it certainly isnt bad, from viewing distance i cannot see any convergence errors.
i will change out the tl071`s and the tl084`s with tle2071 and tle2074 to see if that helps there too.

[tse]

This kinda goes with the original topic. A VDC VIM that never happened. A couple of PCBs were made then the project was abandoned. A pic of the 200MHz neck card prototype.

Scott

[Nashou66]

Scott, I always wondered about the wavy trace lines on some circuit boards. What is their purpose?


Athanasios

[Sparky015]

[dvh99]

i see the prototype had the smb connectors why were they not used instead of the mini rca.
the smb connector with the 4 legs is also commonly available.

[dvh99]

simple question here, is the vim a 3-layer board?

i wonder what is better, getting the contrast signal straight from the opamp mc34084 which i will replace btw or getting them from the testpoints near the rca connectors.

[dvh99]

pic of the vim addon pcb



plus backside, i will order 1 board next week plus all components.

[dvh99]

as you can see on the layout the vdd and vee voltages are connected.
is it better to get rid of them and feed 3*vee plus vdd to the ad834 chips.
or leave them and still feed 3*vee plus vdd to the ad834 chips.

[dvh99]

1 more question, in the schedule of the 03p vim a diode lies on "digital ground", but doing a continuity test from "analog ground" to "digital ground" it says resistance 0.

this confuses me and i need to know what the difference between the groundplanes is cause next week all components will arrive and i will start with the board when the pcb is made.

[tse]

I see you have been working on the PCB layout. It does look better. A couple of things that I noticed, might not be a problem, I've shown on the attached pic. The blue lines are kinda long. HF traces can never be too short. A PCB trace can be 1nH/mm. A 20mm trace at 200MHz can be a significant impedance. The yellow line looks like there might be some clearance issue. Cheap bd houses usually want more than 7 mils.

The schematic from the app note shows 1uF decoupling caps on the multiplier. I don't know of any 1uF caps that have low impedance at 200MHz. I have had good luck with 1000pF COG or NPO type ceramic caps right at the IC power leads with short traces to ground plane. A larger cap a little further away if necessary.

Don't worry about analog and digital ground here. Everything is going to be analog ground. They do have to connect together somewhere perferably somewhere where there are no large high frequency currents.

Scott

[dvh99]

thanks scott for your explanation, the clearance is only shown on the picture but not on the print from the layout.
the pcb is 11*12cm so it will fit nicely on the vim.
the guy who will be making the pcb`s offers an option to add a silver layer to the entire pcb, i read that is a good thing in hf circuits.
i have shortened the blue line on the left with a trace that goes underneath.
i could make the blue lines coming from the transistor wider if that helps lowering impedance
the holes for the connectors are for the smb connectors.
http://www.farnell.com/datasheets/55118.pdf
in the layout you can see 2 holes on the + and -, those are for the 1uf oscons i want to use, do you think placing a 1000pf cap, in parallel with the oscon, from the link would improve things?

dennis

[tse]

The silver is probably a good thing but you shouldn't have to go to the expense. I've done alot of HF bds with only one ounce copper (two ounce is more standard) tinned plating and they have been ok.

I do believe that impedance is determined by length and width with L / W = Z. So wider does decrease Z.

http://www.kemet.com/kemet/web/homepage/kechome.nsf/weben/kemsoft

Very informative free Spice program that shows the impedance of various caps vs frequency. Above a cap's resonant frequency it will actually become inductive. Not what you want decoupling the power rails on HF opamps.

Placing a 1000pF COG ceramic right on pin 6 of the multipier with the other end directly to ground plane would be optimum. The way it is now there is a fairly long trace between pin 6 and the decoupling cap. Undesired inductance in series with cap lowers resonant frequency. I can't say it won't work but why not make it as good as possible?

Scott

[dvh99]

you are absolutely right, i will make the trace wider and i have shortened the trace already it now goes straight to the ad834.
should i leave the oscon there and add the cog ceramic or replace the oscon with a cog ceramic.