What did you buy today? Post your latest purchase!

[PlainName]

Nice. Fabulous if it looks like the one here:

http://www.lathes.co.uk/hardinge/

[dmills]

That is a newer one (Buttons for the speed controls), and I doubt my example will turn up anything like as clean.

We will see next Thursday when the pallet arrives.

[mansaxel]

Hardinge HLV with chucks, steadies, taper turning and a full set of collets.
 
It remains to be seem how much of a strip down and rebuild job it is going to be, not been run in 12 years apparently, so I bet the lube tanks are going to be gnarly. Hopefully nobody used soluble oil in the thing.

Now just waiting for the pallet with 770kg of lathe on it to arrive, then comes the 'sort out a three phase supply' fun.

In Sweden, all houses have 3-phase. If your powering becomes too tedious, there is a home for the lathe...

[dmills]

Oi, get your own, this ones mine

The fun with running these off an inverter is that the fast/slow control is done by switching the motor from two to four pole operation by means of an interlocked contactor, and this can be done while the motor is running... Most inverters get upset by that kind of thing.

Also, selenium rectifier stacks for the carriage drive motor, yea those can fuck right off, easy fix obviously. Interestingly the carriage drive motor does field weakening for high speed mode, not something you expect to see.

[SiliconWizard]

Just received 2 pairs of optical transceivers today to experiment with high-speed optical links.
(Those are bidirectional 1.25Gbps 1x9 transceivers.)

[mansaxel]

Just received 2 pairs of optical transceivers today to experiment with high-speed optical links.
(Those are bidirectional 1.25Gbps 1x9 transceivers.)

I casually put into operation 2 links of 4x25Gbit/s capacity today. (4 wavelengths in one fiber, 1295.56, 1300.05, 1304.59 and 1309.14 nm, adding up to 100Gbit/s, 100GBASE-LR4, IEEE802.3ba-2010)

1.25 Gbit is impressive, and more than most people would need for many different purposes, but high-speed it is not, not anymore. (I run singlemode 10Gbit links at home, so it is common.)

Granted, a coherent 100G system, that is something extra. Not for the faint of heart. 4x25G is much easier.

[dmills]

Meh, got switches in the home office that do that, sort of need them when messing around with ST2110 routing and trying to get the shitshow that results from letting web 'programmers' come up with an interoperability standard for real time doings to actually work properly.

Granted, I am not doing the WDM dance.

Saturating a 100Gb link into a switcher or multiviewer is disturbingly easy with a reasonable number of 4k60 4:4:4 feeds.

Yea, I am NOT a fan of NMOS (Which does NOT refer to the old logic family with no noise margin). 

[boffin]

The postman brought my nice Katsu mini drill press today.

So far, I'm favourably impressed. I'm sure the simple triac based variable speed motor controller can be improved, but it's a surprisingly solid machine. The oversized base is cast iron and the head, cast alloy. In addition to the variable speed, there's a 3 step pulley change allowing it to get up to 8500rpm - An ideal companion for my big Sealey pillar drill when using small drill bits. Something I was worrying about when I ordered it was accuracy, but I can't detect any slack at all between the quill and the head casting and there's no noticeable runout on the little 6mm chuck. No flex between the head and the base either.

The wiring looks ok. It's properly earthed and there are decent insulated crimps and additional fibreglass sleeving where the wires pass the motor. I might put some additional strain relief on the motor wiring though as the motor and pulleys go up and down with the quill.

Overall, really chunky for such a small capacity drill press. The motor is only a 100W brushed, but that ought to be sufficient for small diameter drills - maybe I'll experiment with a brushless one at some point as there's plenty of room.


P.S. The adjustable quill stop seems a bit odd - I think they've put the scale sticker on in the wrong position!

I've had one of these for a while.  Works great for PCBs and 1mm drill bits, lacks torque for the larger stuff >5mm.  I've snapped a couple of belts over the years though, but they're pretty easy to find.  Bought mine from Warren Buffet (riogrande.com)

[mansaxel]

Yea, I am NOT a fan of NMOS (Which does NOT refer to the old logic family with no noise margin).

I sort of look at NMOS and wonder if it isn't the last glimpse of interoperability before the proprietary solution monster eats us all.

The big problem with the business is that it is driven by vendors who have no clue in the Internet standard development model, who are trying to sell complex interdependent networks (more complex than they comprehend) to people who have no idea at all about networks, and are used to paying way too much for special hardware, when all of this, today, can be done on mostly standard gear...

I've got Nevion "control" (the guys who do not believe in multicast routing protocols), and there's LAWO "control" (the solution that can't even control all of their own gear nor interoperate with their own control system) around the corner. I am not impressed with either. And none of them have left the mind model where a facility is only so large as the reach of a piece of Belden coax with blackburst on. ST 2022-7 is completely, utterly incompatible with anything except one studio facility. I've got 4, in a square, with a diagonal added in of several hundred kilometers of wonderful 100G links in between. How do I map this to a very dumb A/B redundancy model?

</rant>

[armandine2]

arrived a day early (amazingly)

Makita DMP181 (10 mins on / 5 mins off) cordless Tyre inflator

[Peter_O]

Have that Maikita too since strange air selling machines have replaced the old free air cans at the gas stations over here. 
Does the job. 

[RAPo]

An analog central multimeter, bust be from the seventies

[dmills]

I've got Nevion "control" (the guys who do not believe in multicast routing protocols), and there's LAWO "control" (the solution that can't even control all of their own gear nor interoperate with their own control system) around the corner. I am not impressed with either.

I thought Nevion was basically a bit of Sony?
 
Lawo is a funny one, I mean Ember+ is basically what happens when you give a German SNMP and tell them to make something almost, but not completely, completely unlike this.... They could SO easily have grafted on a few extensions for discoverability and realtime data and it would have just worked, but no, have you SEEN that spec? It has wonders in it, like the fact that the 'oids' are NOT required to be consistent from one run to the next, it is the text strings you have to match on.... <Spit>. And then you get VSM, classic consultantware.

Trouble is, with GV clearly doing the pivot into being a software company with cloudish aspirations, that don't leave a whole lot of choices, BNCS maybe?

The thing that narks me about the whole NMOS thing is that absolutely NO consideration appears to have gone into doing little things like plugging an IP truck into an IP facility over IP.... You pretty much CANNOT do it without a LOT of forward planning on both ends, and this even applies when using trucks from a few different vendors on one job, it is all SDI and MADI between the islands of IP.
 
</rant>

[SeanB]

Got 3 RS branded Nixie tubes, red glass ones, still new in the RS box. Also a Sharp Compet calculator, though that likely does not work.

[mansaxel]

I've got Nevion "control" (the guys who do not believe in multicast routing protocols), and there's LAWO "control" (the solution that can't even control all of their own gear nor interoperate with their own control system) around the corner. I am not impressed with either.

I thought Nevion was basically a bit of Sony?

They're not Sony yet. They're still on their worryingly static platform with Mellanox switches that only do PTP in Transparent mode.

Lawo is a funny one, I mean Ember+ is basically what happens when you give a German SNMP and tell them to make something almost, but not completely, completely unlike this.... They could SO easily have grafted on a few extensions for discoverability and realtime data and it would have just worked, but no, have you SEEN that spec? It has wonders in it, like the fact that the 'oids' are NOT required to be consistent from one run to the next, it is the text strings you have to match on.... <Spit>. And then you get VSM, classic consultantware.

Actually, we've got a buncha people who actually grok VSM. We did the entire "decouple control room from studio" automation, the thing that people usually "go to IP" for today (which means a big chassis switch and lots of converters  ) on SDI 12 years ago. With VSM as the motor.

Trouble is, with GV clearly doing the pivot into being a software company with cloudish aspirations, that don't leave a whole lot of choices, BNCS maybe?

The thing that narks me about the whole NMOS thing is that absolutely NO consideration appears to have gone into doing little things like plugging an IP truck into an IP facility over IP.... You pretty much CANNOT do it without a LOT of forward planning on both ends, and this even applies when using trucks from a few different vendors on one job, it is all SDI and MADI between the islands of IP.
 
</rant>

Yes, it's a quagmire. I wish that broadcast vendors learned to shut up and not try to build a kingdom that requires perfect obedience in every detail, and then try delivering that turn-key. It never worked before, not even when all was BNC and XLR. I want building blocks (like "audio mixing"  or  "audio source",  "camera signal", etc with good, publicised, standardised interfaces.  Then I want to require them all to tell me they're there, and how to control them. And that they tell that to something that they find using a modern provisioning system, not a decades-old hard-coded address.

And, if they say "mdns", I will get right violent.

With the connect two trucks scenario, it's basically the same as "let's try and merger two companies who both use Net 10" -- which I've tried.  The broadcast world seems hell-bent on reinventing every mistake the corporate net world has made the last 30 years, only with more bandwidth.

[dmills]

MDNS? Yea, that was AMWA/NMOS wanting to do the existing tools thing, completely unsuitable IMHO for this kind of scale (Especially the way NMOS does it).

Mind you AES dropped the ball as well with not putting the minimal amount of metadata needed to successfully decode an AES67 flow into the flow itself... Grumble.

There does seem to be a LOT of reinventing the cockup going around, the control systems are grossly unsuitable (They think in terms of routers still, not controlling endpoints and trunking), the customers don't understand the technology, the standards folks mostly don't understand broadcast, and the SIs are doing their usual trick of over promising and then blaming the vendors. 

I loved that there really does not seem to be a way to build an NMOS configuration in advance of having all the kit on site and connected to the registration server, I mean being able to build the config back at the office before getting to that expensive time on site would make all this shit way too easy would it not?

While SDI had its issues, god knows, SWP08, Quartz and TSL UMD got you control of basically everything, would that the new world was as easy.

[mansaxel]

MDNS? Yea, that was AMWA/NMOS wanting to do the existing tools thing, completely unsuitable IMHO for this kind of scale (Especially the way NMOS does it).

Mind you AES dropped the ball as well with not putting the minimal amount of metadata needed to successfully decode an AES67 flow into the flow itself... Grumble.

I remember having a few discussions with people involved in the AES67 group about how Certain Vendors didn't want things to be too easy. And, as I'd by then built a complete SIP-based registration system (kicking EBU into issuing an updated Tech 3326 in the process) for remote participants, I kinda thought it would work out any way. And, it actually does. But it is very much implementation specific, because someone (I guess both the "Let's run PTP1" people down under and the Italian Town gang) wanted their little walled gardens unthreatened.

There does seem to be a LOT of reinventing the cockup going around, the control systems are grossly unsuitable (They think in terms of routers still, not controlling endpoints and trunking), the customers don't understand the technology, the standards folks mostly don't understand broadcast, and the SIs are doing their usual trick of over promising and then blaming the vendors. 

I loved that there really does not seem to be a way to build an NMOS configuration in advance of having all the kit on site and connected to the registration server, I mean being able to build the config back at the office before getting to that expensive time on site would make all this shit way too easy would it not?

While SDI had its issues, god knows, SWP08, Quartz and TSL UMD got you control of basically everything, would that the new world was as easy.

Basically yes. People want "make TV" look and feel the way they did experience in the 70s, and nothing to change. The "I want a box to be playing the role of router" paradigm is the most annoying part. The network is the router. It is a if not country- so at least site-wide (mostly) non-blocking matrix.  People can be made to understand it, but you have to repeat it over and over.

[dmills]

Amusing to see the PTP V1 Aussie gang and the Italian town gang agree on anything.
Doubly so when you note where the person the PTP V1 gangs main product is named after died...

I despise standards committee bullshit like that, just makes everyone's lives harder.

Still could be worse, SMPTE managed to come up with a PTP profile that actually breaks some standard PTP receive code! The facepalming was allegedly epic. Course the Canadian cousins (Burlington Ontario branch) then went and implemented what SMPTE had cooked up instead of IEEE1588, not that interop was ever much their thing...

Still think SMPTE taking a whole pile of FREE IETF standards, packaging them up with very little additional text and then selling the bloody things was cheeky.

[beanflying]

Completed the circular purchase today when I picked up the new Plasma Cutter  Realized I now need to add an extension to the welding table for cutting on and likely then make a CNC table sometime later 

[DiTBho]

Bought a license for MobaXtem-Professional, 1 user
v21.5 is the best I have ever tried 

And it runs on both Windows XP and Windows 10

[Gixy]

MobaXTerm: this tool has been designed by a friend of mine, in his garage, and now he has clients like NASA!

[PlainName]

I've been trying it out and it's pretty cool, but one of my big usages makes it quite annoying. What's the best route to finding if it's just a config option I've missed or if it can be otherwise sorted?

Specifically, a session starts with the size of the terminal tab. Suppose I use RDP to access some remote machine - I set the session to have a detached tab (it wants to be a free-standing window) and that detached tab is the same size as the current main terminal tab. If I change the size manually, it's never remembered but always starts with the current terminal tab dimensions (this is one of the huge issues with using a browser as an app).

This matters to me because that RDP window will be, say, 1024x768 but another RDP window will want to be, say, 800x600. I've tried fitting to window (just makes the remote the wrong size), smart sizing (just distorts the view), etc. I am tired of having to spend time dragging window corners to get rid of scroll bars.

Further, why should a GUI window want the same dimensions as a text console? And this doesn't even broach the issue of being afraid to resize the main app window because it will affect every session started from then on!

I must be missing a simple option along the lines of "save per-session window size" but I can't find it. If I could I'd spring for a license, no problem

[SiliconWizard]

Just received 2 pairs of optical transceivers today to experiment with high-speed optical links.
(Those are bidirectional 1.25Gbps 1x9 transceivers.)

I casually put into operation 2 links of 4x25Gbit/s capacity today. (4 wavelengths in one fiber, 1295.56, 1300.05, 1304.59 and 1309.14 nm, adding up to 100Gbit/s, 100GBASE-LR4, IEEE802.3ba-2010)

1.25 Gbit is impressive, and more than most people would need for many different purposes, but high-speed it is not, not anymore. (I run singlemode 10Gbit links at home, so it is common.)

Granted, a coherent 100G system, that is something extra. Not for the faint of heart. 4x25G is much easier.

1 Gbsps is indeed considered "high-speed communication" in the digital design context these days. Just a technical term. I thought we were on some electronics forum, not Twitter.

As the word 'experiment' and the bare transceivers suggested, my intent is to design my own links, using FPGA and stuff. Prototype stuff. "Advanced" DIY. Not using existing networks and switches.
While doing that for 1 Gsps is already challenging, it's doable. But I dare you to do that kind of thing with 25 Gbps or higher. This is certainly nothing "casual" ( ), unless you are just using off-the-shelf equipment, which I thought was clearly not the case here, but if it wasn't clear, now it should be. =)

[Martin72]

Simply two boxes, one with 500 LEDs, the other with 6 different colour silicone-isolated leads AWG22, including cableties and shrinking tubes.
For under 20 bucks...

[mansaxel]

1 Gbsps is indeed considered "high-speed communication" in the digital design context these days. Just a technical term. I thought we were on some electronics forum, not Twitter.

As the word 'experiment' and the bare transceivers suggested, my intent is to design my own links, using FPGA and stuff. Prototype stuff. "Advanced" DIY. Not using existing networks and switches.
While doing that for 1 Gsps is already challenging, it's doable. But I dare you to do that kind of thing with 25 Gbps or higher. This is certainly nothing "casual" ( ), unless you are just using off-the-shelf equipment, which I thought was clearly not the case here, but if it wasn't clear, now it should be. =)

To be very explicit, it was the orders-of-magnitude dissonance between what is being casually used and what was described as high speed that bothered me enough to dare writing this. I fully appreciate and admire the challenge of doing it as "advanced DIY", but I felt that the mountain of giants standing on giants that the modern comms industry has built for itself to be able to reach present levels of performance could do with some pointing out.

Simply because we take it for granted.

So, no offense intended.