Startech VID2HDCON2 / QED QCV EA0053 / Koryuu Video conveters

Recently I got a couple of video adapters, so I thought I'd post my quick tests with them. One is a SCART to component converter, the other a S-Video (S-VHS) and Composite to HDMI converter. I also re-tested the Koryuu, which converts S-Video and composite to component. For my uses I'm just interested in using them to get old games consoles working with a modern TV (Philips 50PUS8204/12).

Startech VID2HDCON2

Firstly the Startech VID2HDCON2 S-Video and Composite (and sound) to HDMI converter. This has an S-Video in port, Composite video in (RCA) port, and 2 RCA ports for left and right sound in. Then an HDMI out port.

It takes 5V power, but the jack is smaller than a standard barrel jack. (Mine didn't come with a power supply so I used a USB to barrel jack power with a barrel jack adapter to step it down to the smaller size). Then there's a power switch to turn it on and off.

On the front you get a button to toggle between Composite and S-Video and some LEDs to show you which one is selected and that the device is powered.

While this converter does work okay, it makes the 4:3 ratio consoles stretch to 16:9. Changing the aspect ratio on the TV settings has no effect. So no good to me.

Koryuu

The following tests were using the Koryuu to convert S-Video to Component, then the OSCC to convert to HDMI. The Koryuu has S-Video and Composite inputs on one side, and Component out on the other.

The Koryuu works okay. Have to press input button on Koryuu for a couple of seconds so it changes from red to green light to select S-Video. The Koryuu needs to be turned on before the console(s). Suggest: OSCC on, Koryuu on, change Koryuu to S-Video, change OSCC to YPbPr, turn on console (and select correct switch on S-Video switch if using one).

For sound you need a dual female RCA to 3.5mm stereo male cable. The stereo end goes into the AV2 audio in port of the OSCC. The small switch on the OSCC between this port and the HDMI jack needs to be positioned towards the edge of the case for the jack to be used as AV2 audio in rather than AV1 (SCART) audio out. The Koryuu itself does nothing with sound.

If the TV keeps saying "Video format not supported", putting the TV in standby then turning back on should fix it.

QED QCV EA0053

The QED QCV EA0053 has SCART input on one side and Component out on the other.

When I purchased the QCV it didn't come with a power supply. I found online suggestions that it used a 7.5V DC power supply. This was just mentioned in an online forum, I couldn't find anything more official about the power supply needed. So I tried 7.5V and it does work with that. Possibly it may work with 5V, but I've only tried with 7.5V DC (positive tip). The jack size is standard barrel jack.

Like the Koryuu, I tested this with the component out going into the OSCC. The OSCC does SCART RGB in, so this converter isn't actually needed for this, but I just purchased it because it was cheap and wanted to see how well it worked.

The main issue with the converter is that it doesn't deal with audio at all. So you need to use a SCART pass-thru adapter with audio out into the converter so you can extract the audio. I didn't think the adapter we had would work for this, but it did.

Using RGB SCART into the dumb switchable SCART / RCA adapter into QED SCART to Component converter then to the OSCC makes the sound & video cut out every so often (with PAL GC as source). It does the same without the dumb adapter (but obv. no sound). SCART straight into OSCC has no issue.

Running the component out from the QED QCV straight to the component in on the TV (so bypassing the OSCC) gave no issues. The image was stretched to 16:9, but changing the aspect ratio on the TV settings fixed this. So there seems to be something about the component output from the QCV that the OSCC doesn't like.

When trying this initially, connecting straight to the TV didn't black out but I got 2 bars of weird static effect that slowly move down the screen over the picture. After trying some higher quality audio RCA cables (I don't have long enough high quality component cables, just cheap ones) then I found it actually gave a good quality image.

Using the dumb S-Video to SCART adapter with a S-Video input into the QED SCART to Component converter gives no image or a very dark grayscale image. TV also says "Video format not supported". So the QED QCV does need RGB SCART and won't work with S-Video or Composite (both of which can be transmitted over SCART).

Conclusion

If you don't have an OSCC and want a cheap way to get RGB SCART into a modern TV that supports component, then the QED QCV is a cheap way to do this.

The Startech VID2HDCON2 is not a good way to get S-Video and composite video into a modern TV, at least for stuff that uses a 4:3 aspect ratio, which will be most things you'd want to use it for. The Koryuu works well. If I didn't have the Koryuu, I would probably look at the Kaico OSSC None RGB Add On Board, which seems a bit better featured.

If I didn't have the OSCC (or even if I did but I had the spare cash), I'd look at the RetroTink 5x Pro, which does about everything you need for connecting older games consoles to a modern TV.

Annoying facebook marketplace scammers

I've been looking on Facebook Marketplace for a couple of things I wanted recently. Yesterday I was surprised when I saw a load of listings for one the items I was looking at, all with the same photos but for sale in different locations and by different sellers. There were 3 different listings they had copied and then added multiple copies of each.

This screenshot is actually from today, and there are only a couple of obvious scam items on it

I thought this was quite a weird scam since Facebook marketplace is for face to face sales, and wondered how these scams would work. Well, today I received a reply to a message regarding a listing I had enquired about a week and a half ago. This was before I had seen any of these scam listings, and it wasn't one that had been duplicated. But the reply indicated this was a scam listing as well. Their reply was:

Yes, in our website for sale, limited promotion. If you are interested, please contact me， I can provide you our website， you can learn more about this product in detail. Our products return within 30 days, with a two-year warranty ，free shipping.

So unfortunately it now seems that finding genuine items for sale on Facebook marketplace is very difficult. With eBay listings you can usually easily if its a scam (No Buy it now, but a buy it now price listed in the description, description saying you need to contact teh seller before bidding, etc). For these Facebook marketplace items it seems you need to click on the seller's profile image, and view their other items for sale. If they have a load of other items available and all for the same price, then it indicates a fake scammer account.

I just checked the seller profile on another of the same item I'd been looking at (this one was located further from me, so I hadn't actually messaged them), and sure enough that one is a scam account as well.

Calumet MF6820 Studio Clamp (Cheap Super Clamp)

I purchased a couple of these a few weeks ago, and just got to checking them out today. They are a cheap version of the Manfrotto 035 Super Clamp. I thought I'd do a quick review / info piece on them as there doesn't seem to be much info out there.

In terms of quality, they are what you'd expect - not as good as the real Manfrotto super clamps. I don't have any cheap Chinese ones to compare, but based on my experience with smaller cheap Chinese clamps, I expect the Calumet ones would be significantly better.

The knobs / handle in particular feel a bit more flimsy / cheap compared to the Manfrotto super clamp. On one of my clamps the rubber at the end of the moving part of the jaw is a bit out of position and raised up (I have tried to push it into place, but with no luck - I think it has been glued like this.

The rubber pad on this grip is raised up at this corner

Now, the main point of this post is to cover the use of the 2 screws that they come with, and the threaded holes in the base of the clamp. As there are no instructions or documentation included with the clamp, and no mention of what these are for on the websites of retailers selling the clamp. The point of these is to allow mounting one clamp at a right-angle, attached to the other clamp. So you could have one clamped to the vertical pole of a stand, then the other holding a pole horizontally like a boom arm. Or use double this setup with two stands and the clamps holding a horizontal pole between them to act like a background stand.

The screws push down in the holes within the V of the jaw. They do need a bit of force to push down - use the hex key also included with the clamp (the screws have hex heads). You'll also need to fully open the clamp to more easily get the hex key above them to push them down and tighten them. These then screw into the two threaded holes at either side on the base of the other clamp. Then you do the same with the other clamp so they are both screwed into each other for maximum stability and strength.

The two holes in the V grip part of the clamp the screws need to be pushed down through

Two clamps, one with the two screws pushed through, ready to screw into the threaded holes in the base of the other clamp.

The two clamps mounted together at right-angles using the screws

This combination gives you the equivalent of the Manfrotto 038 Double Super Clamp. But two of the Calumet MF6820 studio clamps are £28 total, while the Manfrotto 038 Double Super Clamp is £54. So quite a saving with the Calumet clamps. However, do note that getting the screws out again from the Calumet clamp is quite a pain.

While the Calumet clamps aren't as good quality as the Manfrotto ones, there are actually some advantages of the Calumet ones:

• They also have a 1/4" female thread in the centre of the base, so you can screw them straight on top of anything with a 1/4" thread. Or you could screw in something like a 1/4" male to 3/8" male adapter then mount something onto the base of the clamp that way (though I would assume you'd generally be better off using a stud with the thread you want).
• You can pull the clamp handle out and then reposition it without tightening / loosening the clamp at all. Useful for repositioning the handle after tightening it down, or if you have the clamp in a position where you can't fully rotate the handle to tighten / loosen the clamp. I believe Manfrotto now makes a super clamp with a ratcheting handle as well, but my old Super clamp doesn't feature this.
• They come with a flat wedge that fits into the V of the clamp for use when you are clamping two flat surfaces together. This neatly stores under the screw used for tightening the stud / spigot in place. However, it does restrict how much you can tighten down this handle when it is stored there. But it's nice that you get the flat wedge included even if the storage solution for it when not in use isn't perfect.

To sum up, initial impressions are that these do appear to be decent, value for money clamps. I would prefer they had better quality to match the Manfrotto super clamps, if that was possible without pushing the price up more than a couple of quid.

Cleaning / Removing Parcel tape / Packing tape residue from cardboard box

tldr; Use a hairdryer to heat tape to allow you to remove it easily leaving minimal residue. Then rub blu-tack on any residue left to remove it.

It's really annoying when you purchase a collectible off ebay, then the seller just tapes it up and posts it off in it's original box rather than placing it inside a separate shipping box. With collectible stuff the box is an important part of the collectible, so to get it covered in large shipping labels and ugly brown parcel tape is a pain.

This recently happened to me, so I've been looking around for solutions to clean the box up and get it looking decent again. The most common suggestion I found was just about using a hairdryer to heat up the tape to allow you to remove it more easily. This did work, though with the amount of tape my package was covered in, it took a long time to get it all off. You just have to heat a section of the tape for half a minute or so, pull that bit up, heat the next few inches, pull that up, etc. until you've eventually got it all off. I also found it didn't work so well where the tape was going over a paper sticker, it was difficult to pull the tape up without ripping the sticker.

After doing this I was still left with a lot of sticky brown residue in various areas where the tape had been. The worst areas were on a couple of paper stickers where the brown residue had just stuck to the stickers and just the plastic part of the parcel tape had lifted up off them.

Most solutions for cleaning up the actual parcel / packing tape residue were to use something like 'Goo Gone'. I have ordered a bottle of this, but my experience with a similar product is that it is quite oily / greasy. While it may make it possible to clean up the tape residue I would expect it would soak into the cardboard and stickers, just giving me a different problem.

A solution I only saw once was a comment from someone suggesting blu-tack (well, actually they suggested the equivalent product from some other country). So I gave this a go, just rubbing the blu-tack on the tape residue and it worked surprisingly well.

Above is actually after cleaning some of the residue from the box, but you can see there is still some residue left on the box, and the sticker is really bad. You can also see my blob of blu-tack I used for the cleaning.

Above image shows partway through cleaning. You have to rub quite a lot with the blu-tack to get the tape residue off. Possibly if you heated the tape residue again if would come off easier? I found that it was particularly stubborn to remove from the paper sticker.

Above shows after cleaning.

Above you can see that this solution unfortunately doesn't work for all cases. While the blu-tack did remove the brown sticky tape residue, it also removed the ink from the printing on the sticker. So I decided to leave this sticker as it is. Maybe I will try some of the goo gone on it when I get that, unless anyone has any better ideas?

New PC (Case)

The past week I've been working on modding an old PC case to move my old / current PC into. The idea being my new PC can then go into the case of my current PC.

The finished PC, in place and powered on

Inside the PC

The PC consists of the motherboard, CPU, M.2 drive, and RAM from my old PC. The CPU Cooler, PSU, etc. from the old PC are being retained for the (real) new PC, and I bought a new cooler and PSU to be used with the old motherboard in this build.

Mesh front and fans

The case I got for fitting the old PC into is an old Gigabyte GZ-X1 Mid Tower case. Like the vast majority of PC cases it is poorly designed, with little thought given to keeping components cool. However, other than this, the design compares very favourably to most other cases.

It has space for quite a few 3.5″ drives, and four 5.25″ bays. In my case I decided to dispense with the 5.25″ bays (at least from an external perspective) and fill the front with fans. This is important for good airflow over any hard drives in the case.

The first thing in terms of the case mod was to dispose of the worse than useless plastic case front. Next job was cutting away the metal front of the case (normally hidden behind the plastic front).

The top 5.25″ bays have metal covers you can just twist to snap off. The lower half of the case front though is just solid sheet metal, except for some breathing holes near the bottom where a 120mm could be mounted behind. I am a 140mm fan guy (The larger the fan, the more air it can move, so lower RPMs, and thus lower noise) so this was no good for me anyway. I just used a dremel to cut away all this section of the case, keeping the same spacing as the top section of the case.

I also had to cut out a section of the bottom part of the 5.25″ bays cage to enable fitting a 120mm fan in the middle, crossing between the bottom (3.5″ bays) and top (5.25″ bays) halves of the case. The case will only fit two 140mm fans and one 120mm fan, hence my use of a 120mm fan here. It also works much better with the spacing available.

I used some wire mesh ~6mm spacing on the front of the case. This doesn't block any airflow to the fans, while preventing you / kids from accidentally / purposefully putting your fingers or anything else into the fans. I managed to secure this with screws going into holes already drilled / punched into the case - they weren't actually threaded but the screws went in okay with a bit of force to get them to tap.

Front view of the case with the wire mesh, fans, and front I/O

Front I/O

Above the 120mm fan in the middle I was left with a space similar to a 3.5″ bay. I thought that would make a good space for some front I/O. I had a USB 3.0 w/ HD audio front panel I had purchased previously. The board from this almost fit, but not quite. I had to cut / grind out a slot in the bottom of the 5.25″ bays cage that the end of the front panel PCB could slot into. I also had to cut out a small section of edge of the 3.5″ bays column to allow for the HD audio header connector to the front panel PCB.

Rear of the front I/O

I had a USB 2.0 rear slot plate, so I unscrewed the sockets from this, to screw them into the front I/O area on the case instead. I also had a USB-C cable to attach. However, after attaching one of the USB 2.0 sockets and the USB-C socket, I found there wasn't enough space for the second USB 2.0 socket (you get 2 sockets on one header). I tried filing down the plastic around the USB-C and one of the USB 2.0 sockets in order to get them to fit one above the other, but without success.

I decided to just go with the one USB 2.0 socket instead, but once I had started assembling the actual PC in the case I found the cable was too short. The other issue was it meant the front I/O ports were very close together, so you'd have trouble plugging multiple things in at one. So in the end I removed the USB 2.0 socket and put the USB-C socket there instead. Annoyingly, this was after when I'd already cut the mesh for where I had been trying to position the sockets. So the mesh around the I/O area is not quite as neat as it should be.

While the I/O ports are right up against the mesh, this doesn't seem to cause any problems. I can plug USB devices and and audio cables in no problem. When testing the audio I did notice it was a bit buzzy, but with the front I/O board loose (so not touching any metal), and only the HD audio connected (USB headers disconnected) it was still the same. Now I have the PC set up at home I also get slight buzz via my USB audio interface, so my guess is it's down to the cheap PSU (Cylon 700W) I'm using.

Power button

For the power button I had the reset switch (I think from the same case, the power switch from the case I was using with the board for my new PC). Then a plastic power button. The plastic power button is meant to attach into the plastic case front (I don't think it was actually from the Gigabyte case I was using, but another case). Since my modified case doesn't have any way of attaching the button, I had to make something up.

I hot glued the push part of the switch onto the rear of the plastic button, so pushing the plastic button will push the switch. For attaching the switch to the case I bent a rear slot plate / cover, so it would form an S shape (with the top of the S being quite long). I then epoxied the bottom of this onto the top of the reset switch.

The plastic button I was using had some bits that poked out alongside the switch. When I glued the bracket onto the top of the switch I found the glue squeezed out onto this plastic, so I had to insert a piece of paper between the switch and the plastic (only thing thin enough) to prevent them being glued together (which would have meant you couldn't depress the switch). Once the epoxy had dried, I managed to get most of the paper out, but there was some left in the area where the actual glue was. Thankfully, pressing the button depressed the switch and ripped the paper through the middle, so this did work to prevent them being stuck together. It was a bit rough when pressing the button, so I sprayed some WD40 on the glued area and this sorted that out.

For attaching the switch to the case I could now put a screw through a hole in the side of the lowest 5.25″ bay and into a hole in the bracket. I was thinking I would need to drill out a hole in the bracket and use a screw with a nut on the other side, as I couldn't fit a screw into the bracket's existing hole. But Will tried it with a more wood screw type screw, and this went in okay.

The mounting solution for the power button, viewed from below

The button I'm using did have a hole to insert a power LED into (so the button lights up when the PC is on). But I don't like coloured lights near my PC (bad for colour accuracy) and I can tell if it's on by if the fans are spinning or not. So I left this out.

External SATA

Since I am still going to be using this as my current PC until I get the new one up and running, I needed some way of connecting hot-swap HDs. I need one for my 2.5″ SATA SSD that I store my work on, and one for my 3.5″ SATA HDs used for daily backups.

External SATA cables

So for this I just ran a couple of long SATA cables and a power cable with a molex to SATA then SATA to 2x SATA power adapter attached. While not quite as convenient as a trayless backplane, it should still be pretty easy to connect / disconnect the drives as needed. And unlike a backplane, it doesn't matter the physical size of the disk being plugged in.

Cable Management

In terms of cable management, it's not something this case really provides. And using a non-modular PSU like the one I'm using, just worsens the problem. The PCI-e power cables and 2nd SATA / Molex cable that I wasn't using, I just tucked up in the corner of the case. The other cables I tried to neaten up / tie together as best I could.

Top down view

Spare cables pushed into the corner

Front I/O and fan cables (inc. extender cable for the power switch cable)

The fan cables I used a single header to four fans adapter. The Gigabyte fan (which I think comes from this case) has about a 5cm fan cable, and the Noctua fan only has around a 10cm cable. So the use of the adapter cable was more for giving enough extension to allow plugging the fans in rather than the motherboard not having enough fan headers.

Molex power for the fan splitter cable

USB-C cable and the 3.5″ hard-drive mounted

Drive bays

For adding in 3.5″ hard drives, with the motherboard in the case I found that I couldn't insert the drive into the top slots. To be able to do this I'd need to remove the RAM and the front fan on the CPU cooler. So this case design would not be suitable if you needed to keep the machine on when swapping out a drive (assuming you had to use those drive slots).

As I'm only currently putting a single HD in it, I just put the drive in a lower slot. Even if I did have to use those slots that are blocked by the RAM / cooler fan, it wouldn't be a big problem as the only time I would need to do that would be when installing the drive.

Hotswap bays are best for allowing easy changing of drives, but I'm not really a fan of them for permanently installed drives. This is because they tend to have very poor airflow over the drives, so the drives get very hot and are more likely to die.

Case orientation

I initially placed the motherboard in the case for a quick test, before I had removed the CPU heatsink I intended on keeping for use with the new PC. And I found that actually it did fit and the case lid would close up okay with this larger heatsink (Noctua NH-D15S). So I am now thinking I will do another similar mod on one of these cases for my new PC, rather than using the 5U Server chassis I was using previously / intending on using for the new PC.

The only negative of this is that the case lid / side panel has a raised area above the CPU. I don't know what the reason for this is, but it will prevent stacking one machine on top of another when using them in desktop / rack orientation.

Raised bump on the side panel over the CPU area

The case is around 4U and on my shelf two of these stacked would fit nicely. I think what I would need to end up doing is running the bottom one with the side / top panel removed. I did look to see if there was such a thing as a PCI-e to PCI-e link device for connecting 2 PCs together, as the top PC could also have it's bottom panel removed then a cable used to link the 2 machines together. However, it seems that while such a thing (though not just a simple cable, it would need a logic board) would be possible, it doesn't exist.

The reason I prefer rack mount orientation is because it puts the weight of the heavy CPU cooler pushing down on the CPU (as it should be), rather than trying to bend the motherboard as it would be with the case in tower orientation. Similarly, with a GPU (though I don't have a decent one currently), rack orientation puts the weight of the card pushing down into the slot rather than pulling down on the slot as it would be in a tower case.

For watercooling I can understand a tower case. Similarly, for basic GPUs and stock / cheap CPU coolers, a tower case orientation may well make sense. But for weighty air coolers (both on CPU and GPU), rack / desktop orientation is much more sensible.

For my new PC I'm thinking rather than the wire mesh front I might go for some wood, then drill lots of holes in it for airflow. But we will have to see how that goes.

PS/2 mouse not working via KVM with Windows XP

I've been trying to get a couple of old PCs and servers running recently. I have (or at least plan to have) them connected up to a KVM switch so I don't have to faff around plugging and unplugging the keyboard / video / mouse when I want to switch to a different machine.

But the Pentium 4 era Windows XP machine doesn't seem to accept the PS/2 mouse when it is plugged in via the KVM. Plug the same mouse directly into the PS/2 port on the PC and it is fine, just not via the KVM. I initially thought it was a dodgy cable (I checked plugged one of the servers into the same ports on the KVM, and the mouse worked fine, so I knew it wasn't the KVM). So I purchased another couple of cables, which arrived today - but still the same issue.

I've now found the cause / fix for this, so I thought I'd post it here. In the BIOS Advanced > Peripheral Configuration > Mouse Contoller had to be set to Enabled.

The BIOS on the particular machine gives a helpful description of the meaning of this setting. You can set to Disabled to disable the PS/2 mouse port and free up an interupt input (IRQ). Mine was set to Auto, which is meant to free the interupt input if no mouse is plugged in, but use it for the mouse if one is detected. So it must be that due to the KVM, the mouse prescence wasn't being detected and so the PS/2 mouse port was being disabled. Changing the setting to Enabled (always enable to PS/2 port for mouse) fixed the issue.

RTL9210 NVMe USB3.1 vs SATA SSD

Quite a while back I bought a "USB3.1 Type-C Enclosure" for use with my Corsair Force MP510 512GB NVMe drive. After I purchased it, I found that it was only going at / below USB 3.0 speeds, but with a different NVMe drive installed, it would work at 3.1 speed (or at least above 3.0 speeds). I'd not managed to find any fix for this, so I've just been using it since then at the USB 3.0 speeds.

I've been working lately with my web development site files on my home PC's internal SSD, which isn't great as it means I can't take them to work (or anywhere else) with me. Previously I had been using a Sandisk Extreme USB 3.0 memory stick, but I'd been having quite a few issues with that (it's quite a few years old now), hence why I had moved them to the internal SSD. This weekend I was trying to tidy up my web development files, so I thought I would move them to the MP510 NVMe drive in the USB 3.1 caddy, which I currently just use for various VM VHDs. It wouldn't be as fast as the internal SSD, but at least it would be portable and faster / more reliable than the USB stick.

I thought I would take a look again at seeing if I could fix the issue with the speed, and another problem it has, which is dropping connection when on a USB 3.1 socket. First thing I tried was updating the ASMedia USB drivers for my motherboard to the latest version. Unfortunately that made no difference.

The board for the NVMe to USB caddy uses a Realtek RTL9210 chipset. I managed to find a firmware update to 1.23.15.111620 for the chip. (Finding the firmware update for this and the latest USB drivers for my motherboard was a pain - it seems you have to find them on third party sites and the manufacturers don't release them directly. Yes, motherboard manufacturers do have driver updates on their sites, but these tend to be old ones.) The update went fine, after doing this I ran CrystalDisk Mark on the drive, then unplugged the drive to move it from the USB 3.0 socket it was curently plugged into, to a USB 3.1 socket.

But the disk didn't show up, and when I went into Disk Management it said the disk needed initialising. I tried rebooting, unplugging and re-plugging a few times, tried it in a different PC, but no luck. So I took the drive out of the caddy, in order to try it bare. Unfortunately my PC only has one M.2 slot, and this is used for the boot drive. So I couldn't test it in there.

I looked through all my adapters, hoping for an NVMe riser cable or PCI-e to NVMe riser cable, but the closest I had was a "NFHK N-M201 Ver. 2.0" PCI-e to NVMe M.2 card. I plugged this into a x4 speed slot on the motherboard and thankfully the drive did show up okay.

MP510 NVMe drive via PCI-e to NVMe adapter card

I then tried a PCIe x1 - PCIe x16 riser, with the cable and x16 slot board outside the case (to allow for easy plugging / unplugging of the drive without opening the case up). Then plugged the PCIe to NVMe board into this. This worked, but speeds weren't great.

MP510 NVMe drive via PCI-e 1x to 16x riser and PCI-e to NVMe adapter card

I had issues with my PC refusing to boot, but eventually tracked it down to my use of a USB 2.0 extension cable that only had a single 4 pin USB connected into it, leaving 4 bare pins. As I had been messing about inside my PC I think I must've moved this cable so the pins were touching something metal and shorting out. I couldn't find any electrical tape that wasn't rubbish so I just covered up the bare pins with some selotape to resolve the issue.

Next I tried a simple PCIe x1 - x1 riser. For some reason they had made it with the end of the slot blocked off (so you could only physically fit a x1 card into it). So I had to dremel out the end of it to let the x4 size PCIe to NVMe adapter fit. But this gave the same results as on the x1 - x16 riser.

MP510 NVMe drive via PCI-e 1x to 1x riser and PCI-e to NVMe adapter card

I then tried a Crucial MX500 1TB SATA SSD I'd purchased recently and not used, with a SATA to USB 3.0 adapter. But this just crashed Crystal Disk Mark. So then I tried it connected via SATA, and the speeds were actually better (at least for sequential reads and writes) than the NVMe drive.

MX500 SATA drive via SATA

After that I tried again with the NVME to USB caddy in a USB 3.1 port, and it now worked. But results weren't any better.

MP510 NVMe drive via RTL9210 NVMe to USB in USB 3.1 socket

MP510 NVMe drive via RTL9210 NVMe to USB in USB 3.0 socket

So it looks like my way forward is just using the SATA SSD connected via SATA. I'm not sure why the NVMe drive in the PCI-e adapter is so slow, particularly on the x1 risers. The only thing I can think of is if the N-M201 PCIe to NVMe adapter card was actually PCIe 2.0 speed rather than 3.0 - but I'm not sure how that would be possible when it's nothing more than traces from the PCI-e slot to the M.2 slot and an LED.

The MP510 drive attached to a PCIe - NVMe adapter, inserted into a PCIe 1x - 1x riser, the empty PCIe 1x - 16x riser, and MX500 connected via SATA.