That, and Rigol do not do component level repair either. My DSA815-TG lost the 900 MHz- 1.5GHz range literally 2 days after the warranty expired. "Oh, yes, you can send it in for repair. We do module level repair only, ~€ 900 ex VAT plus shipping cost". Sure, do you think I am a lunatic? The lower two ranges (looks like it has 3 LO ranges) work ok. Now that I have found myself a nice HP 8595E SA I might give the Rigol some trouleshooting time.
Removal of part identification numbers not only conflicts with the right to repair, but it also supports the use of low-grade parts, black market trafficking of parts and counterfeit part manufacturing. It'll take forever to get the legislators involved to change their policy though.
@@pa1wbu One more reason to rapidly start working on domestic production of high-end electronics in both US and EU with a new mindset: longevity, sustainability first. I'd rather pay 2K for something I'd never have to replace than be forced by manufacturers to buy new ones at a fraction of that because it cannot be fixed, upgraded, supported anymore. I have a few heirloom tools that are still as good as the day they were forged; for everything I build myself that is the benchmark.
Yeah, I have 'tons' of classical T&M from HP, Tektronix etc. Still very much ok and with excellent service manuals. Ok.. they were not €2k new. More like 8-40k$ each when new in the '80s and '90s 😉
Seen a number of SA's with blown front ends in my career. All because some incompetent user connected a transmitter directly to connector without suitable attenuation. Nice repair.
Excellent job at explaining the signal path and the different components and their functions. Maybe having your mouse in a different color would make following even easier. Keep up the great content.
Great video, i always learn a new idea for repair from you. I didn't know you could count turns on smd inductors, then looking for matches in the pcb is obvious, but only once you pointed it out! Thanks again for sharing
Excellent repair as always. I also thought that the PIN diode polarity was odd and would have expected two pairs back to back. However, looking at the picture you found on line (what luck - many thanks to whoever posted it) the PIN diode polarity is arranged three to one (????) Regards, David
Congratulations! 😊 You have done a great job on this Rigol. I have this very same Spectrum Analyzer, so this may help me in the feature. You never know how much this could help me for feature reference. It could have happened to any of us.
3:35 Looks like a few components in the tracking generator stage have been replaced too, does the TG work? They also need the flux washing off with some IPA!
Would the carbon left behind be a problem, ie, most sand it all off. Is 60db floor normal for cheap SA, I guess I'm used to high end from Signal Path - lol -.
I'd really like to know what blew the front end like that. Maybe it was attached to an external antenna and got a healthy electrical discharge ? Is there a DC bias voltage on the pin diodes ? That could explain the difference - not sure why they'd use pin diodes it all they're doing is providing protection. I just read some of the comments further down, surely someone didn't transmit into the SA, did they ? 🤔
You never know. Shit happens. Perhaps, someone made a big mistake. Then opened the instrument, saw the damage, said "oh, shit", put it back together not caring much about the screws going into the right holes, and sold it on ebay.
I have Rigol DSA815-TG spectrum analyzer which measures 25dbuV as noise floor. The noise level does not change irrespective of grid connections, and even if I test it with multiple sources. What do you make of 25dbuV? Is it too much to call the AFE to be damaged? I don't think so. But I think there must be something to bring that noise floor down to
nice repair. not so difficult, but good work here. I can say i will not solder this pin diodes at all. If too much power will be applied to the inpey this diodes will not help)) 1 or two switches will blown up again. What type of 1st mixer in this schematics is?
are you sure they must be PIN diodes? for protection? I dont think that is right. Also 10nF cap in the 1.5GHz signal path does not seem right either (except for the input AC coupling cap which is large)
I am almost sure about the diodes. I am not sure at all about the capacitor. However the input cap measures about 900nF (in circuit). The one in the termination path after the switch also measures about 900nF (again, in circuit). What effect would you expect if the capacitor is too low or too high?
PIN diodes are often used as a limiter in RF circuits - when in the off state they present very low capacitance to the circuit, they have linear resistance to high frequency RF so do not introduce harmonics or intermodulation products, they switch states very quickly, and, when conducting, have a resistance that is inversely proportional to the signal level.
@@feedback-loop The cap would form a 1st order high pass filter, so if the value is too low the frequency response will roll off at the lower end of the frequency range. If it's too high, it might bump the noise floor up a little, but not entirely sure what else it might do.
Great job. Crazy how test equipment used to have service manuals and come with schematics. Now they go as far as to laser off the part ids.
That, and Rigol do not do component level repair either. My DSA815-TG lost the 900 MHz- 1.5GHz range literally 2 days after the warranty expired. "Oh, yes, you can send it in for repair. We do module level repair only, ~€ 900 ex VAT plus shipping cost". Sure, do you think I am a lunatic?
The lower two ranges (looks like it has 3 LO ranges) work ok.
Now that I have found myself a nice HP 8595E SA I might give the Rigol some trouleshooting time.
Removal of part identification numbers not only conflicts with the right to repair, but it also supports the use of low-grade parts, black market trafficking of parts and counterfeit part manufacturing. It'll take forever to get the legislators involved to change their policy though.
Getting that sorted with manufacturers in China would be a Herculean task at best I'm afraid :-(
@@pa1wbu One more reason to rapidly start working on domestic production of high-end electronics in both US and EU with a new mindset: longevity, sustainability first. I'd rather pay 2K for something I'd never have to replace than be forced by manufacturers to buy new ones at a fraction of that because it cannot be fixed, upgraded, supported anymore. I have a few heirloom tools that are still as good as the day they were forged; for everything I build myself that is the benchmark.
Yeah, I have 'tons' of classical T&M from HP, Tektronix etc. Still very much ok and with excellent service manuals. Ok.. they were not €2k new. More like 8-40k$ each when new in the '80s and '90s 😉
Perhaps the loss in amplitude from the diodes can be calibrated out. I'm sure the instrument would need cal after such a front end repair regardless.
Seen a number of SA's with blown front ends in my career. All because some incompetent user connected a transmitter directly to connector without suitable attenuation. Nice repair.
Excellent job at explaining the signal path and the different components and their functions. Maybe having your mouse in a different color would make following even easier. Keep up the great content.
Maybe use what Shahriar @ TSP uses: a mouse pointer with a circle around it? Makes it much easier to follow.
Yes, I need to look for some software to change the cursor and, perhaps, draw on top of PDFs and photos. Do you know of some for Ubuntu?
I'm sorry, I don't. The days I ran FreeBSD and Linux on my desktops are long ago. But I would assume there is a solution for Linux!
Still the best channel on youtube! I started watching this channel when you were dissecting crystal oscillators. Keep at it please!
Very nice repair! There must be a calibration procedure to settle the input gain offset you have. 👍
Great video, i always learn a new idea for repair from you. I didn't know you could count turns on smd inductors, then looking for matches in the pcb is obvious, but only once you pointed it out! Thanks again for sharing
Excellent repair as always. I also thought that the PIN diode polarity was odd and would have expected two pairs back to back. However, looking at the picture you found on line (what luck - many thanks to whoever posted it) the PIN diode polarity is arranged three to one (????) Regards, David
Yeah, not sure why this is so, but the same way in both units.
Отличный формат, увлекательно, и не затянуто, всё по делу.
Congratulations! 😊
You have done a great job on this Rigol.
I have this very same Spectrum Analyzer, so this may help me in the feature. You never know how much this could help me for feature reference. It could have happened to any of us.
Great job!
Self-disassembly technology by Rigol ? :)
Great job, man!
Thanks for sharing your videos
A good pragmatic repair as always
3:35 Looks like a few components in the tracking generator stage have been replaced too, does the TG work? They also need the flux washing off with some IPA!
Yes, the TG seems to work fine
Would the carbon left behind be a problem, ie, most sand it all off.
Is 60db floor normal for cheap SA, I guess I'm used to high end from Signal Path - lol -.
Great work....cheers !
A very good repair.
I'd really like to know what blew the front end like that. Maybe it was attached to an external antenna and got a healthy electrical discharge ? Is there a DC bias voltage on the pin diodes ? That could explain the difference - not sure why they'd use pin diodes it all they're doing is providing protection.
I just read some of the comments further down, surely someone didn't transmit into the SA, did they ? 🤔
You never know. Shit happens. Perhaps, someone made a big mistake. Then opened the instrument, saw the damage, said "oh, shit", put it back together not caring much about the screws going into the right holes, and sold it on ebay.
99.9999999% of the time.. Someone directly attaches a transmitter and sends watts of RF power right into the SA.
wow Mr. Fdbcklp. Great!
I have Rigol DSA815-TG spectrum analyzer which measures 25dbuV as noise floor. The noise level does not change irrespective of grid connections, and even if I test it with multiple sources. What do you make of 25dbuV? Is it too much to call the AFE to be damaged? I don't think so. But I think there must be something to bring that noise floor down to
nice repair. not so difficult, but good work here. I can say i will not solder this pin diodes at all. If too much power will be applied to the inpey this diodes will not help)) 1 or two switches will blown up again. What type of 1st mixer in this schematics is?
Seems like HMC213A: www.analog.com/media/en/technical-documentation/data-sheets/hmc213a.pdf
I thought I saw damage to Q203 just to the right in the next box, but it might have been the light.
How did you know 18A means 150 ohms and 57X is 38.3 ohms?
This is a standard code. eepower.com/resistor-guide/resistor-standards-and-codes/resistor-smd-code/
I did not know about the EIA96 markings thanks!
👍
3.03 It’s hard to see, but Q203 seems to be damaged
I don't think so, just some flux residue. It might have been replaced though.
Cross your fingers. Is the generator working as well?
Yes
a other great repair more test equipement saved from the junk pile
are you sure they must be PIN diodes? for protection? I dont think that is right. Also 10nF cap in the 1.5GHz signal path does not seem right either (except for the input AC coupling cap which is large)
I am almost sure about the diodes. I am not sure at all about the capacitor. However the input cap measures about 900nF (in circuit). The one in the termination path after the switch also measures about 900nF (again, in circuit). What effect would you expect if the capacitor is too low or too high?
PIN diodes are often used as a limiter in RF circuits - when in the off state they present very low capacitance to the circuit, they have linear resistance to high frequency RF so do not introduce harmonics or intermodulation products, they switch states very quickly, and, when conducting, have a resistance that is inversely proportional to the signal level.
@@feedback-loop The cap would form a 1st order high pass filter, so if the value is too low the frequency response will roll off at the lower end of the frequency range. If it's too high, it might bump the noise floor up a little, but not entirely sure what else it might do.
Interesting concepts for trouble shooting and parts replacement selection. Enjoyed, thanks.