ITS1099 Intersil NPN Transistor datasheet ?

[floobydust]

Another approach is seeing if the input impedance has gone low due to a failed JFET - gate shorted to D/S. I think R1 surge-resistor carbon 6k2/R3 51k1 should have no real voltage drop across it, but it will if a shorted JFET is loading down that line or trying to put -30V to the D/S. Same for REF Hi and REF Lo, look for a shorted protection JFET as if this unit was overloaded on ACV.

I would try pulling out/lifting a gate leg on a few JFETs to test them, and change to using a sine-wave or straight DC and recording voltages up to Q27.
Or do comparison ohmmeter readings between the two boards. It's not easy troubleshooting because the JFETS have common connections, so narrowing it down to the exact component based on measurements might burn up one's patience.

[RichardM]

I thought this maybe a easy repair for a hobbyist given I have a fully functional second unit. I enjoy the learning experience. It must be a delicately balanced circuit for the active filter module.

I pulled Q18, 25 and 32 and they test ok in the curve tracer. They are by no means great matches, a bit like many transitiors. I wonder how well matched they need to be ?

I feel like I have probed resistances between both boards without finding any huge variations. Diode tests of all the Jetstar and transistors also appear consistent between both filter modules and no shorts have jumped out at me.

I did also test Q27 in circuit with my curve tracer and in the filter that does not work correctly the transistor appears down in gain at low base current. Would this be critical ?

I will keep probing as per suggestions. Thankyou.

[RichardM]

I have tested the dual suoer beta transister Q27 (IT1099) in circuit and attached the results. I normally pull transistors out for testing but I am concious of pulling pads from the old circuit boards. Q27+ and Q27- are as shown in the schematics above.

There is a noticeable difference between the good and bad active filter module measurements but whether this makes a difference of is affected by components downstream is above my knowledge

Any comments always welcome.

Richard 

[m k]

Have you tried those other filters?

[RichardM]

I tried changing filter modes and it made no difference to the general behaviour. I am unsure where and what to probe when changing filter modes ? Opamp outputs ? The timing somewhere ?

[m k]

It seems that Ohms is not using default fast filter but can't confirm.
I have a recollection that high ohms is unfiltered but it can be from somewhere else.

Store 1 Filter uses your suspicious bypass route.
Toggling the filter button is toggling between slow and fast filter.

If your volts and amps are fine then everything after filter selection is also fine.
Everything before filter selection can also be fine.
So Q27 is fine and Q18 is not blocking.

Testing Q18 side selections is easy if you have Ext Ref connectors installed.
Just connect something, select and see the correct result.
There input is multiplexed, all 3 are measured in one pass.

For filter selections you have test points C1/2, C5/6 and C15.
You should see how signal come and go when you switch between filters.
If needed you can poke U6 pins and check how path selection happens.

Use DC volts mode.

My only Ohms controller is disconnected and "far away" so can't check how it behaves.
Can't find anything rejecting filters with it either.
So when you're done with DC volts compare results with Ohms.

[RichardM]

"Store 1 Filter uses your suspicious bypass route. Toggling the filter button is toggling between slow and fast filter."

Quick update as I continue to probe.

Whilst in DC Volts the volatge readings are fine for F (slow filter), F0 (fast filter), F2 & F3. The DC voltage reading is NOT correct in F1 (filter bypass) mode.

The external reference works as expexted on the good unit but the bad unit reads random numbers like 0.0001321E^6

Richard

[RichardM]

I think the only thing left for me to do is pull all the jfets and check them one by one.

I have probed resistances all over the good filter board and bad filter board and they compare well
I have replaced U5 and that made no difference
I have replaced Q6 and Q12, no difference.
I have checked all transistors and diodes in circuit and all appear fine
Volatges rails also ok

My latest probe test was to set the unit in DC volts, probe Q12 collector whilst setting filter to F1 (bypass). The good unit sees the volatge move from -30V to +4.8V. The bad unit sees the volateg move from -30V to around +0.4V ?

[Kleinstein]

When the fitler is set to bypass, the colector of Q12 should go to the votlage of the buffer amplifier output, normally the voltage the DMM reads.
The filter channels should be similar: -30 V when off, and the buffer output, when activated.

[RichardM]

When the fitler is set to bypass, the colector of Q12 should go to the votlage of the buffer amplifier output, normally the voltage the DMM reads.
The filter channels should be similar: -30 V when off, and the buffer output, when activated.

OK, thankyou . So my probing was with +5V DC on the inputs which explains the 4.8V on Q12 collector ? What could be the reason for only 0.4V (not 5V approx) on Q12 in the bad filter module ?

FYI, all the Jfets tested fine on the curve tracer. All 15 of them !! I note thy are market SF 2799.

Richard

[magic]

Possibly reverse breakdown of one of the JFETs that are driven to -30V or the signal coming from the left side of the board is 0.4V in the first place. Trace where 0.4V is coming for, check for voltage drops across all those series resistors.

[RichardM]

Thanks, I will follow the 0.4V.

I am finding it hard to reconcile why the Ohms Module readings are so random on all the filter settings F0 to F3 and DC volts is fine unless the filter is set to bypass F1.

Cheers

[m k]

Well, at least you have some JFET tracer pictures for future reference.

Your Ext Ref reading indicates that signal is over voltage protected or very small signal / very big Ext Ref.
I assume those colored schematics greens are from no Ext Ref situation.
How are those CR1 and CR2 zeners?

Check also some timings.
When Ext Ref is not used Q3 is always on, it's clearly fine then but with Ext Ref something happens.
Q3 can be lagging but Q0 can't, it would finally be ok, but it can be connected to that Ext Ref situation and give that high omhs all the time, obviously only if filter is bypass all the time.
So first check that there are no extra lines activated when filter bypass is on and Ext Ref is off, no timing needed, U1 74174 clear is not used, so outputs are steadily what they are.
Then check that Q3 is not lagging by putting Ext Ref on and checking how Q3 changes with rising edge of pin 9.

[RichardM]

I haven’t pulled CR1 or CR2 but the voltages either side are the same for the working module Andy the bad module. A DMM diode check also indicates they are ok.

I will try to probe a small you suggest. Should I do this in DCV or OHMS mode ?

Thanks again

[RichardM]

There is a bit of variation in the J086/2086 JfET traces, particularly the last  one ?

[Kleinstein]

The JFETs are uses as switches and not as linear amplifiers. The JFET traces look OK and show working jFETs. However this part is not relevant for use as a switch. For the use as a switch the main relevant part is getting low gate leakage even with a relatively high gate voltage (e.g. up to 50 V in this circuit). A possible failure mode is a reduce reduced break-down voltage for the gain.

The other property would be the DS leakage when the fet should be off, though this should cause less of a problem.

A point to check may be the voltage at the collector of Q10 when the FETs are removed  (may need a ground connection (like 100 K to ground) for the follower input if all 6 fets for the fitler settings are removed).

[m k]

I haven’t pulled CR1 or CR2 but the voltages either side are the same for the working module Andy the bad module. A DMM diode check also indicates they are ok.

I will try to probe a small you suggest. Should I do this in DCV or OHMS mode ?

Thanks again

Use DCV with meaningful signal, 10V is up side of X1 range, including meaningful Ext Ref.
Start with standard fast filter and Ext Ref off.
Put 1st scope channel to C1/2.
You should see your signal all the time, nothing is changing the signal path so you can also check how those bipolars are operating.
You should also write down some truth tables, there are quite many things to remember.
Then poke Ext Ref on.
Now your 1st scope channel should see 3 different levels, that didn't happen last time.
You can now use the 2nd scope channel and check where those missing ones are, leave the 1st channel to be a sanity check, assuming there still are sane Ext Ref levels.
You can poke Ext Ref and Filter buttons freely, this is a pure digital tech on/off switching situation.
Signal path selection can also be different than expected, so check selections all the way back to U1.

Don't forget that the problem can also be a trace and not a component.

[RichardM]

OK, that was a bit of work. Checked all the timings and the good and bad unit are the same, see attached.

Where does that leave me ? One of the 15 Jfets are leaky ?

Richard   

[m k]

I had no idea what exactly happens during Ext Ref, now we know.
Good to also know that it's all rational.
Since you've left signal values out I assume there were too many actual levels.

Switch to Ohms, select 1M range and measure a bit less than 10M resistor.
Use default fast filter and check it actually is that and not bypass, like in Ext Ref situation.

Your RT6 should pulse +/-18,5V.
(manual picture 602-3. Timing)

Your earlier Error 6 is hinting that incoming 18,5V is dropping too fast but accurate DCV is pointing elsewhere.

Some DC 10V range fast filter errors would be nice, are there any?

[RichardM]

All the voltages are 0 to +5V in the tables. V=10V at inputs and Vext=15V at tear.

I am going to check Q1 to Q4 transistors now.

All the DC voltage DMM readings are fine except when when Vext is on or Bypass from memory.

[m k]

There's practically only C15 left.

Have you checked the continuity of the bypass route?

[RichardM]

When I select bypass F1 in DCV mode the input RT6 measures at 28.7V on the broken filter module and large voltage drops are on R1 and R3 wrt TP1 The good module just reads the 10V input value. This is really confusing me

[RichardM]

There's practically only C15 left.

Have you checked the continuity of the bypass route?

C15!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Measures 2.2nf 66 Ohm ESR but < 100 Ohm resistance across it.

It didn't look suspicious when I measured it in circuit from memory. Rookie error I suspect, although I had thought these poly film caps are generally not the source of concern. Lesson learned I suppose.

My 8502a is now working in DVC & OHMS mode, now to run it through some tests. Sure to be another error thrown up The poor old filter module has been pulled in and out a thousand times and almost every component desoldered and resoldered/replaced. It has lost a few pads in the process.

Thank you all for your assistance and hanging in there with a beginner.

Richard

[m k]

Practice is always good.

It's also easily misleading that this module has logic Vcc actually -15V and so R(TP1) is -20V.

If you one day return to that area you could check how Ohms mode is using filters.
I read that filters are always free to use.
Now it seems that Ohms is bypassing all the time.