Could someone create a topic with the final fix for the older SDG1000 hardware? I'm not sure what needs to be changed.
It depends what is problem and what is HW etc.
1. Square wave and same glitches in both channel and IF there is factory installed around 40k R87&R82 (PECL comparator hyster setting resistors)
This can reduce changing PECL comparator hysteresis window - IF it is too narrow.
Perhaps best and most "safe" compromise is change R87 and R82 to 20k (19-21k but so that they are equal). More radical values are 12-15k BUT if use these it need carefully test they do not affect too much Duty % accuracy. Also too wide hysteresis window (smaller resistance) may work still good with 50% duty but some problems may rise if use other than 50% duty if hysteresis is too wide. It is easy understand if look "theory" how this circuit work.
2. There may be situation where CH1 and CH2 have big difference in square (jitter/glitch) In this case Bad Channel may help with small capacitor what filters internal sinewave what is coming to PECL. (there is 50 ohm line terminator resistor just before PECL. Small capacitor parallel with this resistor may help.)
3. If there is wrong amplifier in CH2 AND if it produce parasitic oscillations using some levels and specially mismatched output loading. This amplifier need change. (this is quite rare problem - in some lot factory have installed other manufacturer pin compatible amplifier. Right amplifier in CH2 is EL5166. and afaik, only this.
If there is this problem, it affects ALL waveforms in CH2. (knowing this may help diagnose without opening case)
Ch1 do NOT have this problem!
First need carefully investigate what is problem exactly (or if there is not problem) and then think what is solution just for this case. If gliches level is well under 10% of square level and if 1kHz timing jitter (period time error related to ideal time ) is around 0.1% or less it tells it is ok and no need repair. (If want improve it better than specs, it is other case)
Always, also after repair, Square wave time jitter is quite high. (but around inside specifications)
If need Square wave, and if need less jitter and if frequency is under some hundred kiloherz and if need much much less jitter below 10kHz Square wave. Just do it using Pulse function. Nearly it can recommend to use Square wave only if some reason Pulse function can not use for producing needes square wave. But with higher frequencies (over 5MHz) Pulse can not use. And Square jitter is less than Pulse jitter but I do not remember now numbers. It becouse Pulse have always 8ns system jitter independent of frequency (exept with these "golden" frequencies what can produce without 8ns timing jitter for keep (short time average) frequency right)
Thumb rule: Use Pulse function for Square wave.
1kHz Square wave have 1000000ns cycle time. If timing jitter is 0.1% from cycle it is still 1000ns.
1kHz Square wave made with Pulse function have only this around 8ns jitter.
And what best, Pulse do not have glitches before rising edge and falling edge.
Pulse is made directly by main DAC.
Square is made by PECL comparator from internal sinewave (made by DAC,s of course)
But, it is good to know that becouse this, with higher frequencies, Square wave time jitter is very low (and also with high frequencies clitches level is very low or no clitches)
These some problems have been mainly in last year and partially 1Q2013.
Perhaps later these all are ok from factory.
Also example SDG1010 what I have normally get from factory and produced 1Q2013 did not have any problems.
And as can see in last phasenoise test with unit just as from factory box, they have very good Sinewave if look phasenoise/noise floor sidebands from carrier.
Home
Help
Search
Login
Register
