WizardTim YouTube Channel

[WizardTim]

This thread is about the WizardTim YouTube channel

New videos are posted here as replies, discussions in this thread about the videos are also welcome.




"Taking a look at the construction and inner workings of a common SMT ceramic LCC package quartz crystal oscillator from Daishinku Corporation (KDS) manufactured in the late 1990s."

I’ve tried making videos in the past to give back to the YouTube community that’s always giving although they didn’t really turn out any good mostly because they’re not something I would actually watch. This video however was driven by my actual interest about something I wanted to find out so I think it might also interest to some of you.

Full disclosure I’m shilling myself here in an attempt to get the video out of the YouTube algorithm’s pit of obscurity as I doubt it’s going to serve my video if I just sit back and twiddle my thumbs. If you enjoyed it I would appreciate your likes as offerings to the algorithm and if you think something might become of my “fantastic” production quality consider subscribing to my video RSS feed, or the bell thing to play YouTube’s notification lottery.

Also, I’m interested in your opinion of the actual device, I’m not a silicon design engineer nor a quartz mechanical engineer so I’m sure I’ve missed or got something wrong.

[thm_w]

Way more in depth than I was expecting, great.

So you are saying they use a high frequency crystal and then divide it down, Eg 40MHz -> 10MHz. Is that due to small size of the crystal used?

Here is a PLL style unit but its not cheap ($10): https://www.silabs.com/documents/public/data-sheets/si540-datasheet.pdf

[WizardTim]

I suspect they use an 80 MHz crystal with a divide by 2 (current similar models from KDS list f0 as 80 MHz). However, if this were true, I would expect the quartz strip to be less than 0.05 mm thick but it’s 0.10 mm thick alluding to it being a lower frequency (around 16 to 25 MHz) however I’m not sure if the quartz frequency constant works the same way as conventional round disks as with strip resonators.

PLLs are much more complex and thus a lot more expensive to implement than a simple divider and they tend to have limitations especially in phase jitter and start up performance so I would assume KDS would avoid using one as much as possible, I’m wishing I powered and probed it without the lid, would have made it a lot easier.

For AU$10.45 that Si540 could be an interesting teardown there might be some RF wizardry going on to get the 1.5 GHz performance.

[thm_w]

New videos:

[thm_w]

[WizardTim]

Thanks for embedding the latest videos here, I’ve changed the thread title to ‘WizardTim YouTube Channel’ instead of ‘SMT Crystal Oscillator Teardown’ and will embed any future videos here as I upload them in case anyone wants to watch this thread instead of betting on the magic YouTube notifications lottery.

Also, some behind the scenes info; I wanted to release 3 special videos this month about a current topic and while I did get some of the key scenes recorded and the intro party edited for 1 of them as they’re somewhat expensive and involved videos I’ve decided not to rush them and to post-pone them to ‘later this year’ (Valve Time), they won’t be topical anymore but I want them to be good videos more than I wanted to release them on time for the Algorithm™. Until then I’ve got a couple other videos in various states I need to finish and post (but they’re nowhere near as interesting).

Spoiler/teaser: The 3 specials relate to electronic devices that were used during and in the aftermath of an event 10 years ago in March of 2011. (picture related)

[Bud]

Thanks for the interesting video but audio part could be better, signal/noise ratio is poor, i had to crank up the volume to max.

[WizardTim]

Is that just the SMT crystal video with the bad audio or a recent one? After making that video I went out and bought a proper audio interface and condenser microphone, while it doesn’t make my computers any quieter the noise floor in all the recent videos should be vastly lower.

[Shock]

Great videos, there were a few quiet ones. Google the best post production settings for youtube audio if you haven't already. You cant edit without uploading again sadly.

But here is what you can do, since your channel is new and only has 4k views and a few videos if you are thinking about making it more episodic you can reupload them and edit the titles such as.

WT #0001 - Bench PSU to USB Banana Board
WT #0002 - SMT Crystal Oscillator Teardown

When you reupload videos you can mention (reupload) in the first part of the title or comments and mention you had to improve the audio. A few months later you edit it out and noone gets upset, magic.

[Ed.Kloonk]

Happy to help with audio cleanup if you aren't on a tight time budget and prepared to send the file within the size of gmail's file size limit. Even if that means including a reduced video res.

[WizardTim]

I’ve had a read up on what YouTube expects me to upload, previously I was just normalizing peaks to -1dB all calling it good enough, I see now I should be doing a perceived loudness equalization to about -10 LUFS to keep it more consistent for YouTube to then normalize it again (and probably speaking louder in the first place).
I’ll do this in future but I won’t bother with the existing videos, the really bad ones I did with my old gaming headset microphone are lost causes anyway, they’re just super noisy and normalization won’t help them, they already have a fair bit of noise reduction applied.
I don’t think I will put numbers in the title, I’d rather have the extra few characters before YouTube wraps/hides the end of the title. I have however given each video a reference ID eg. ‘WT-YT-V-9’ so you can search for episodic numbers that way, but I’ve mostly done that for me naming files locally during production.

Thanks Ed.Kloonk for the offer but after having a read I now understand what to do differently so there’s no need.

[WizardTim]

New Video:

"Teardown and repair of a modern cost optimised consumer grade 'frosted candle' style LED bulb from OSRAM and in-depth analysis of the failed component. This light is advertised as 5.5 W 470 lm with a lifetime of "up to" 15,000 hours and costs at the time of writing AU$4.90. It is estimated this light had approximately 2,000 - 4,000 operating hours before failing."

[Shock]

Nice failure analysis, hopefully it's a low failure rate. Not a big fan of Osram for no reason other than a few have died on me prior to me moving to led lighting.

As long as the leds don't fail cascading one faulty globe could service many others which is cool.

[WizardTim]

New video:

"Teardown of a 20 kV 5 mA high voltage diode (2CL77) in a DO-312 package manufactured by GETAI Electronic Devices Co. Ltd. (广州市格泰电子器件有限) under the brand HGVT Semiconductor."



This video was actually filmed in June and it's only now I've gotten around to editing and voicing it, I've got several other videos filmed but not edited or voiced and plenty of half filmed or just ideas of videos however I have too many other hobbies and full time jobs to get those done as quickly as I would like but they will be done eventually!

Last month (August) I bought some new LED studio lights (replacing homemade ones), a new smartphone camera (replacing an 8 year old one) and a new desktop computer (replacing a 10 year old one) so once I finish my back log of unedited videos I should be able to start on 4K HDR 60 FPS videos.

[WizardTim]

New video:

"Taking a look at a suspicious TO-3 2N3055 BJT transistor acquired from an eBay seller for cheap that is allegedly is manufactured by STmicroelectronics."

[WizardTim]

New video (#15):
“Repairing a broken Simpson (Electrolux) washing machine from 2010 with a faulty SMPS (Switch Mode Power Supply) followed by an in-depth failure analysis of the failed component and speculating as to possible failure modes.”

[WizardTim]

New video (#12):

"Taking a quick look at a failed HPM XL632 PIR sensor from the late 2000's and analysing how it failed followed by an in depth teardown of the pyro-electric sensor package (KDS9 5E1), as well as an IRA-S210ST01 manufactured by Murata and comparing it to a ZTP-148SRC1 thermopile manufactured by Thermometrics (Amphenol)."




It's been awhile since the last video but a lot has happened, I've been working on several other videos but I keep increasing the scope of them, I originally recorded the first 10 minutes of this PIR sensor teardown in 2021 just as a test of my new smartphone and some hacked software for it but decided to keep and later edit the footage as a quick filler video, but I wasn't happy with it and abandoned it for several months, upon getting the new camera I revisited it and filmed the rest of the video adding enough to make it worth watching.

This is also the first 4K video which I think turned out quite well, although I was considering doing 4K HDR 60 fps previously I now understand HDR/WCG is not even slightly ready for YouTube or really any platform or device except dedicated closed ecosystems, so this video is SDR in BT.709 although I recorded some parts in BT.2020 & AdobeRGB it was just too difficult and I couldn't be bothered as <<1% of devices even support it properly. I also kept with 30 fps due to recording hardware and file size limitations, although for those types of videos 60 fps doesn't really make that much difference, I'd rather spend my mbps on contrast and sharpness which the new camera has in buckets. I will however note I still have a lot of footage from the old Nokia Lumia 1020 in 1080p before I even bought studio lights or a colour balance card, this footage I can't re-record so the videos I have planned will still have varying image quality or be 1080p.

[WizardTim]

New video (#21):

"Troubleshooting and repairing a broken PIR (Passive Infra-Red) motion sensor sold under the ROBUS brand (RU360-01) and then taking a closer look at the failure mode and discussing the robustness of X class safety capacitors and a teardown of the failed capacitor and speculation about the root cause of the failure."



This is a follow up/continuation of the previously posted video #12. I'm quite happy with how this one turned out as it was pretty straight forward to make, there wasn't much research needed or complex camera/editing work like the other videos I've been working on.

[Ed.Kloonk]

New video (#21):

"Troubleshooting and repairing a broken PIR (Passive Infra-Red) motion sensor

Thanks for the vid. I have a box full of these and similar dead soldiers. Might set about repairing a few in a group if that's all that it is.

I don't know either if it's power fluctuations or summer heat or moisture. Have had a number fail after a big storm. And yes fail means silly behaviour.

Having considered your video, I too suspect moisture. It's very hard with that model to use the rubber seals on the inlet ports when you've got a bunch of grapes coming out of the wall.

 

[Ed.Kloonk]

Had a look at one of mine today. Mine are Matelec brand FMS-50000

https://www.matelecaustralia.com.au

Except for the re-brand and model sticker on the wire cover plate, these are the same right down to even the same shitty Jimson caps.

The two yellow caps were around the wrong way so I wasn't able to read numbers until I removed them. I tested both with two different MM and got almost exactly the same readings you got with your faulty ones!

Put in the Vishay ones and indeed the detector and relay now work better. 

I half thought about modifying on for a 12vdc input (for a location that has only that) but looks like the circuit and relay is 24v. Oh well.

Might be worth mentioning that this Matelec model is supposed to have a 3 year warranty. There is a slightly different size 'premium' model in the range that is supposed to have a 5 year warranty.

Will put the repaired one back in service and reno the 'replacement' one there before it fails. Will be interesting to watch how long it lasts.

Thanks again for the vid. One less thing around here to be grouchy about.

[Ed.Kloonk]

So I went to change a 'newer' sensor which has been turning the carport into a disco during the night. 18 months old. It's been totally respun. Just about the only part that is interchangeable is the AC connector block. 

How I found out, I thought I'd cheat and swap the back casing to avoid having to deal with fitting those outside springs. Opened it up and Houston, we have a problem.

They have changed the board shape, removed the fuse, used a bridge rectifier instead of all those little diodes everywhere, less chips and the plastic mount for the tri-sensor is different and the red LED is missing.

Boards and housings are incompatible because of a added rib where the fuse used to be and the pots have been moved so the knob stems dont line up. 

Image 1 left board is older one with old caps replaced.
Image 2 boards are flipped.
Image 3 New board has different sensor mount but I think the sensors themselves might be the same.

[WizardTim]

Thanks for the follow up and I'm glad to hear the repair worked for you.

Very interesting to see just how much they've managed to cost cut it by, I'm quite impressive they've managed to go from 3 separate PIR ICs to just a single one and they ditched IC5 which I think is an unmarked MCU. But that new PCB feels rather weird, it's almost like they gave a different company one of the old PCBs and told them to recreate it for as little BOM cost as possible as none of the layout or artwork is re-used, even looks like a different PCB EDA but the design is clearly a clone. Also strange they re-did the mold for the plastic sensor standoff, has to be a completely new manufacturer. The external appearance of the 'premium' MATelec FMS-12080 is identical to my ROBUS unit, while the MS-50000 'economy' version is quite different, amusingly the 'economy' MS-50000 is only $10 cheaper than my ROBUS while the 'premium' FMS-12080 is TWICE the price?!

The blue AC PSU cap is probably still the failure mode in the new economy design so hopefully it's still an easy fix.