Tools of the Trade.

[BillO]

I've never used the Rigol. I know Dave Jones over at EEVBLOG seems to like the build quality of the Rigols over the Siglents, but likes the Siglent UIs and layout better. I just picked up a Siglent SDS1204X-E (actually a SDS1104X-E with upgrade license keys) as well to use as my main scope. Just getting to know my way around it, but it seems to have all the features I'll likely need for a while and is fairly easy to use.

[BigDumbDinosaur]

GARTHWILSON wrote:

I wonder how this Hameg [Edit: Rigol] is...

A friend has a Rigol DS2202 and has commented favorably about it. It’s on my short list of scopes to replace my ailing H-P 1725A. The Rigol unit doesn't have quite the bandwidth of the H-P, but has a large (for a scope) display and quite a few useful features.

Quote:

About the Epson LX-350 printer I mentioned above: I never got an answer about the question I emailed through their site about the Esc codes...

Why didn't you ask me about this? I have that info here.

[BillO]

For the first timer this might be good value for money: Hantek Scope

100MHz, AWG and protocol decoding. Nice feature set.

It seems to get relatively decent reviews too, I mean keeping that price in mind.

Edit: So, that one I linked to does not come with the AWG enabled, but all the hardware is installed. The front ends are also capable of 150MHz and apparently it can be easily hacked to full DSO2D15 capability.

https://www.eevblog.com/forum/testgear/ ... e-dso2x1x/

Quote:

Cost-effective economy oscilloscope, 150MHz Bandwidth, 1GSa/s, 8M memory depth; with 1CH 25MHz waveform generator, support arbitrary waveform output; 14 kinds of trigger modes, standard with 5 kinds of serial protocol triggers and decodes; 32 kinds of auto measurements with statistics; 3-digit digital voltage meter and 6-digit hardware frequency indicator functions; 2 sets of DVM; Abundant SCPI remote command control.

Not too shabby for a sub $200 piece of kit.

[GARTHWILSON]

GARTHWILSON wrote:

About the Epson LX-350 printer I mentioned above: <...> so I ordered one. I should be here in a week.

I should give an update. It did arrive last week, and I'm very pleased with it. Print quality is excellent for a 9-pin, and it's faster than any I've seen before, maybe a little smaller too. It's the cheapest of the fourteen (!) models of dot-matrix impact printers that Epson still offers. They have 24-pin ones too. They seem to understand that this market isn't going away, and they've picked up the slack where other manufacturers have stopped making them. I got this one brand new on eBay for $217. It has parallel (which I use), RS-232, and USB interfaces.

[akohlbecker]

GARTHWILSON wrote:

GARTHWILSON wrote:

About the Epson LX-350 printer I mentioned above: <...> so I ordered one. I should be here in a week.

I should give an update. It did arrive last week, and I'm very pleased with it. Print quality is excellent for a 9-pin, and it's faster than any I've seen before, maybe a little smaller too. It's the cheapest of the fourteen (!) models of dot-matrix impact printers that Epson still offers. They have 24-pin ones too. They seem to understand that this market isn't going away, and they've picked up the slack where other manufacturers have stopped making them. I got this one brand new on eBay for $217. It has parallel (which I use), RS-232, and USB interfaces.

Do you use it for anything in particular, compared to, say, a laser printer?
Would be nice to build a parallel port interface and driver for my SBC, wondering how hard that would be

[GARTHWILSON]

akohlbecker wrote:

GARTHWILSON wrote:

GARTHWILSON wrote:

About the Epson LX-350 printer I mentioned above:  <...> so I ordered one.  I should be here in a week.

I should give an update.  It did arrive last week, and I'm very pleased with it.  Print quality is excellent for a 9-pin, and it's faster than any I've seen before, maybe a little smaller too.  It's the cheapest of the fourteen (!) models of dot-matrix impact printers that Epson still offers.  They have 24-pin ones too.  They seem to understand that this market isn't going away, and they've picked up the slack where other manufacturers have stopped making them.  I got this one brand new on eBay for $217.  It has parallel (which I use), RS-232, and USB interfaces.

Do you use it for anything in particular, compared to, say, a laser printer?
Would be nice to build a parallel port interface and driver for my SBC, wondering how hard that would be

In addition to what I wrote in my post on the previous page, I use it most for printing source code, where I want continuous-feed fanfold paper.  With sheet feed, page breaks often come at the worst possible places, in program structures for example.  When our daughter-in-law was majoring in computer science, she complained that the printer she had access to in the dorms at school caused this problem, and she ended up cutting the top or bottom margin off each page and taping them together for a continuous printout to spread out for proofreading a section of code that may be four or eight or twelve pages for example.

With the workbench computer, I sometimes want to print tables or dumps, sometimes only a line or two at a time and I want to see it before continuing, which means I don't want it ejecting the page like laser printers or inkjet printers do.  It also needs to be text only, so the computer doesn't have to figure out which dots to turn on and feed everything as graphics.

I have a little bit about these printers' pinouts and signaling protocols in the 6502 primer's circuit potpourri page, at http://wilsonminesco.com/6502primer/potpourri.html#LPT, and you can see my workbench computer's printer port circuit just above the middle of the page at http://wilsonminesco.com/6502primer/IO_ICs.html, along with the many other things I use VIA1 for.  Because of all the pin multiplexing and port sharing I do on that first VIA, the 8-bit data had to get split up between VIA ports.  The software handles it.

For another printer to print .pdf files, web pages, etc., I've given up trying to keep an inkjet or laser printer working.  We've had several, and they've all been major headaches.  So when I have something like that I need printed, I email it the local UPS Store which has some Kinko's-type services and have them print it for me, and walk up there to get it.

[akohlbecker]

GARTHWILSON wrote:

akohlbecker wrote:

GARTHWILSON wrote:

GARTHWILSON wrote:

About the Epson LX-350 printer I mentioned above: <...> so I ordered one. I should be here in a week.

I should give an update. It did arrive last week, and I'm very pleased with it. Print quality is excellent for a 9-pin, and it's faster than any I've seen before, maybe a little smaller too. It's the cheapest of the fourteen (!) models of dot-matrix impact printers that Epson still offers. They have 24-pin ones too. They seem to understand that this market isn't going away, and they've picked up the slack where other manufacturers have stopped making them. I got this one brand new on eBay for $217. It has parallel (which I use), RS-232, and USB interfaces.

Do you use it for anything in particular, compared to, say, a laser printer?
Would be nice to build a parallel port interface and driver for my SBC, wondering how hard that would be

In addition to what I wrote in my post on the previous page, I use it most for printing source code, where I want continuous-feed fanfold paper. With sheet feed, page breaks often come at the worst possible places, in program structures for example. When our daughter-in-law was majoring in computer science, she complained that the printer she had access to in the dorms at school caused this problem, and she ended up cutting the top or bottom margin off each page and taping them together for a continuous printout to spread out for proofreading a section of code that may be four or eight or twelve pages for example.

With the workbench computer, I sometimes want to print tables or dumps, sometimes only a line or two at a time and I want to see it before continuing, which means I don't want it ejecting the page like laser printers or inkjet printers do. It also needs to be text only, so the computer doesn't have to figure out which dots to turn on and feed everything as graphics.

I have a little bit about these printers' pinouts and signaling protocols in the 6502 primer's circuit potpourri page, at http://wilsonminesco.com/6502primer/potpourri.html#LPT, and you can see my workbench computer's printer port circuit just above the middle of the page at http://wilsonminesco.com/6502primer/IO_ICs.html, along with the many other things I use VIA1 for. Because of all the pin multiplexing and port sharing I do on that first VIA, the 8-bit data had to get split up between VIA ports. The software handles it.

For another printer to print .pdf files, web pages, etc., I've given up trying to keep an inkjet or laser printer working. We've had several, and they've all been major headaches. So when I have something like that I need printed, I email it the local UPS Store which has some Kinko's-type services and have them print it for me, and walk up there to get it.

Very informative, thanks! I'm now looking at getting a second-hand dot-matrix printer to experiment with

[BigDumbDinosaur]

akohlbecker wrote:

Very informative, thanks! I'm now looking at getting a second-hand dot-matrix printer to experiment with

I've got several dot matrix units in my office that are “relics” from when I used to do professional vertical application programming. The one that gets the most exercise is an Okidata Microline (ML)395 that I bought as a refurb in 1996 and fitted with a pull tractor. It’s still going strong and actually does pretty decent graphics...if you’re not in a hurry.

I've never been a fan of Epson dot matrix printers. They don't do well in commercial applications when printing multipart forms—the mechanism isn't robust enough for the job. The Okidata Microline units are nearly indestructible and depending on the model, very fast for that type of printer. I've got two clients using these units to print multipart invoices and statements, as well as greenbar reports, in high volumes. About all they need is a periodic ribbon replacement and after 10 years or so of continuous use, replacement of the print head, which is a five minute job.

[BigDumbDinosaur]

GARTHWILSON wrote:

I have a little bit about these printers' pinouts and signaling protocols in the 6502 primer's circuit potpourri page, at http://wilsonminesco.com/6502primer/potpourri.html#LPT...

From that...

Quote:

The lower two lines, data and strobe, are computer outputs to the printer, while the upper two, busy and acknowledge, are computer inputs from the printer. You do need to pay attention to the busy line, but I've never paid attention to the acknowledge line.

That is at odds with the original Centronics interface, which did not have BUSY. BUSY, along with some other “features,” was added to the Centronics interface by Epson under pressure from IBM, originally in a misguided effort to prevent non-IBM printers from working with a PC. The real Centronics interface uses /ACK as the handshake, along with /STROBE. The parallel port on a PC, regardless of age, is an aberration, not a true Centronics port.

BUSY is not entirely trustworthy as a handshaking line. When asserted, it indicates the printer’s buffer is full or the printer is off-line. It doesn’t mean that the most recent byte was accepted for processing. That is the role of /ACK. Also note that BUSY is high-true, but /ACK is low-true. There is a reason for that...in handshaking, low-true is usually considered the more reliable arrangement due to limitations of TTL totem-pole outputs.

Furthermore, if the port’s driver uses interrupts, use of BUSY as the handshake may cause timing violations that can lead to data corruption. This is because BUSY is asserted almost immediately when /STROBE is driven low. It may not be a problem with a low-speed computer, but will definitely arise on a faster system. Use of /ACK doesn’t open this door, since /ACK is not asserted until well after /STROBE has been toggled and the datum has been accepted by the printer.

Many years ago, I concocted a Centronics interface for my Commodore 128 through the user port, which is wired to CIA #2. The handshaking was /STROBE and /ACK, and I also rigged up the interface so when the computer went through reset it would toggle /RESET on the printer ([pin 31). I wired it in a way so I could reset the printer independently of the computer. This project got published in Transactor.

[GARTHWILSON]

The reason I didn't use ACK\ is that my CA1, CA2, CB1, and CB2 VIA inputs were already dedicated to other things.  They're the only ones that let you do interrupt-on-change, and the ACK\ pulse may be as short as 500ns, much too short for polling.  [Edit, correction: It's the Strobe\ input that has to be at least that long.  The ACK\ is around 12µs, with the falling BUSY edge approximately centered in those 12µs.]  I've never used interrupts in printing, only polling.  Doing it the way I have, I have never had any problem.  Printing with my routines worked on first try, and has never had a single hiccup in 32 years.

I do have an Okidata printer too, but it doesn't use the same IBM437 character set.  I've never had to print multi-part forms.  Edit, 8/2/22: When re-arranging things in the garage yesterday, I realized I also had a couple of wide-carriage Epson printers I had forgotten about.  Maybe I didn't need to spend the money on a new printer!

[floobydust]

BigDumbDinosaur wrote:

GARTHWILSON wrote:

I have a little bit about these printers' pinouts and signaling protocols in the 6502 primer's circuit potpourri page, at http://wilsonminesco.com/6502primer/potpourri.html#LPT...

From that...

Quote:

The lower two lines, data and strobe, are computer outputs to the printer, while the upper two, busy and acknowledge, are computer inputs from the printer. You do need to pay attention to the busy line, but I've never paid attention to the acknowledge line.

That is at odds with the original Centronics interface, which did not have BUSY. BUSY, along with some other “features,” was added to the Centronics interface by Epson under pressure from IBM, originally in a misguided effort to prevent non-IBM printers from working with a PC. The real Centronics interface uses /ACK as the handshake, along with /STROBE. The parallel port on a PC, regardless of age, is an aberration, not a true Centronics port.

BUSY is not entirely trustworthy as a handshaking line. When asserted, it indicates the printer’s buffer is full or the printer is off-line. It doesn’t mean that the most recent byte was accepted for processing. That is the role of /ACK. Also note that BUSY is high-true, but /ACK is low-true. There is a reason for that...in handshaking, low-true is usually considered the more reliable arrangement due to limitations of TTL totem-pole outputs.

Furthermore, if the port’s driver uses interrupts, use of BUSY as the handshake may cause timing violations that can lead to data corruption. This is because BUSY is asserted almost immediately when /STROBE is driven low. It may not be a problem with a low-speed computer, but will definitely arise on a faster system. Use of /ACK doesn’t open this door, since /ACK is not asserted until well after /STROBE has been toggled and the datum has been accepted by the printer.

Many years ago, I concocted a Centronics interface for my Commodore 128 through the user port, which is wired to CIA #2. The handshaking was /STROBE and /ACK, and I also rigged up the interface so when the computer went through reset it would toggle /RESET on the printer ([pin 31). I wired it in a way so I could reset the printer independently of the computer. This project got published in Transactor.

I guess my "Big Blue blood" started this reply, reference the modified Centronics port on the original IBM PC. Going further back in time, pretty much all IBM I/O devices, regardless if they were punch card, tape, disk or printer had a "ready" indicator lamp. When illuminated, the device was ready to accept a command, when out, the device was considered to be busy. Busy being a tape drive searching for the requested data block, or disk drive searching for the requested sector, etc. The same applied to printers, such as the 1403, which the N1 version could print 1100 lines per minute of 132 characters wide and could do a high speed skip of 70 inches per second (announced in 1959 and withdrawn in 1999).

The original IBM PC had a projected life of approximately 250K units.... so much for that guess! In any case, realize that the IBM PC was basically designed by more traditional IBMers, hence the fact that the original IBM PC used parity memory, which all IBM compute platforms did. So I'm not quite sure that it was IBM trying to pressure Epson to make a one-off printer (hence proprietary), versus IBM engineers stating they had to have a BUSY signal in their own mind to be inline with traditional IBM I/O devices and the interactions with the controlling device. I suppose it's anyone's guess.... but then again, as time went on and the PC industry leaped at an alarming rate, IBM did create some one-off bits and such to minimize other players getting in the sand box. But when you publish the entire set of hardware schematics in a complete manner and the entire BIOS listing in the Technical Reference Manual, it's really inviting someone to go off and build their own, as it was all "off the shelf" parts of the day.

Here we are 4+ decades later still talking about it... who knew?

PS - I also wrote code to drive a parallel port printer from my C64 back in the 80's... I bought a Star Micronics dot matrix printer... even as a employee, couldn't afford/justify the employee discounted price on the IBM-logoed Epson printer. Oddly, I still have an IBM ProPrinter that's basically NOS, but I'm certain it needs a new ribbon!

[BillO]

Just an update .. don't want to leave anyone with incorrect information.

1) The UNI-T UTD2102CEL I discussed earlier does indeed have an XY mode. It's just buried in the wrong menu under "Display" rather than "Acquire" where the rest of the world has it. This unit seems to be not available any more, replaced by the UTD2102CEX. Cannot find anything different in the claimed specifications, but it costs less than the CEL did all those years ago. Not that you should care. My advice is to steer clear of UNI-T. They have ZERO support.

2) That nice little UNI-T AWG I was talking about, that a few days ago was selling for $100? It's the UTG962. Well, it seems to have gotten some attention and now is selling for over $182 over on AliExpress. Again, while it seems like a nice little instrument and has recently received great reviews you need to consider it totally unsupported and without warranty.

On a bright note I have to give a thumb's up to KORAD. I have had their KA3305P 3 channel programmable linear bench power supply for about 6 years now. It has worked very well considering it was a fraction of the price of big name units. There are a few quirks, but nothing that impairs usability in any meaningful way. Recently it stopped communicating with my e-bench computer. I sent a note to KORAD and they have engaged right away with the assurance they would help with the problem even though the warranty has expired. The have assigned a single point of contact and things a progressing as well as you could expect. Nice. I will follow up in a separate TotT thread when thing get further along. Even DJ of EEVblog reviewed a single channel version from KORAD and gave it a thumb's up, although given it was not a Keysight or R&S he had to hold his nose and wear gloves to do it. But in reality, most of us do not have $27,000 to spend on a bench PS (if you do, I don't want to hear about it (jk)) but the ~$350 I paid for the KORAD was well worth it.

[Windfall]

BillO wrote:

Even DJ of EEVblog reviewed a single channel version from KORAD and gave it a thumb's up, although given it was not a Keysight or R&S he had to hold his nose and wear gloves to do it. But in reality, most of us do not have $27,000 to spend on a bench PS (if you do, I don't want to hear about it (jk))

I would. At $27,000 it probably has a robot hand that comes out and makes all your connections for you, so you can sit back and direct it telepathically while you drink your coke. I would want to know about it. Not buy it. But know about it. What does '(jk)' mean BTW ?

[BigEd]

(Hi JK - jk means "joking")

[Dr Jefyll]

BigDumbDinosaur wrote:

BUSY is not entirely trustworthy as a handshaking line. When asserted, it indicates the printer’s buffer is full or the printer is off-line. It doesn’t mean that the most recent byte was accepted for processing.

My experience has been the same as Garth's, BDD; ie, never a hiccup ignoring /ACK and using BUSY instead. I'm not convinced I need /ACK to tell me "the most recent byte was accepted for processing." In practical terms, how is that different from knowing the printer is online and has room in its buffer? The only thing my code (Z80 code, in this case) wants to know is, "can I send a byte now or must I wait?"

Quote:

BUSY is asserted almost immediately when /STROBE is driven low.

True, but I'm not connecting the dots on how this constitutes a problem. Are you worried about a high-speed computer reacting too promptly to an interrupt caused by BUSY being asserted?

As for the limitations of TTL totem-pole outputs, I agree there's some real asymmetry there. It's more likely that noise will result in a false high being momentarily perceived on the receiving end than a false low -- IOW, a low is more trustworthy. And I think we also agree it's far more destructive for the host to momentarily perceive a false "go ahead" from the printer than it is to momentary perceive a false "keep waiting." It's more important to minimize the risk of a false go-ahead. And ... isn't that how BUSY is in fact defined? Maybe it's clearer to think of that signal as /Go-Ahead.

-- Jeff [edits]