Tachometer design

[Steve GarnettParticipant @stevegarnett62550]

I think that the project is heading that way, but it hasn’t really got there yet. I think it will, though. Hopefully the discussion has been worthwhile, regardless.

[Ian S CParticipant @iansc]

How about a mechanical system with flyweights connected to an indicator needle,might be a bit watch making,but I might try it.Ian S C

[Sub MandrelParticipant @submandrel]

Out of interest, if people want a practical layout, how about the one I use for my mini-lathe and mill? It fits in this neat transparent blue box about 2″ by 5″. It uses a hall effect switch taht has three wires (ground, 5v and signal) and all that’s onthe machine is the stationary sensor and a small but powerfull magnet, 5mm dia by 1.5mm thick, – separation distance not critical. The magnet is so powerful; it isn’t even glued on! The photos are a bit grainy as I didn’t use flash.

The pics show the speed range this can cope with, and also showcase the greater than 100:1 speed range of a mini-lathe with roller bearings fitted (I never use the ‘high’ range it goes scary fast!).

 

 

[Sub MandrelParticipant @submandrel]

Another thought. With AVRs it would be easy to design a standard “ME Module” with a range of IO ports and programmable over a simple serial link (or USB via an adapter module) by having what is called a “boot loader”.

 

If everyone putting together projects standardised on the same module, chips with the bootloader could be sold cheaply, and the programmes could be downloaded for free from here.

 

With a suitable choice of chip, it would be possible to apply it to all sorts of applications – from clocks and tachos, through DRO readouts and stepper motor controllers, to brushless motor ESCs and process controllers.

 

A standard design would cut costs to the quick, and make it very easy to develop collaborative projects.

 

It’s likely that many hobbysists would prefer to go the PIC route, but if there’s interest in going AVR, I’m up for designing a module and bootloader system.

 

[Steve GarnettParticipant @stevegarnett62550]

Nothing wrong with having more than one similar project going, I would have thought. As far as I’m aware there are basic bootloaders available for both AVRs and PICs, and I would have thought that modifying one of these might well be an option – presumably that’s along the lines you were thinking? I haven’t really looked at the chip options to determine which ones might be suitable, and really, that’s the only issue I have with this approach; you’ve almost got to determine in advance what you might want to connect, and because your bootloader is determined only for a specific device type, the moment somebody wants to do a project that exceeds the capabilities of it, they’re a bit stuffed.

 

Mind you, to be fair I don’t think that this is a very serious issue – even with PICs the amount of stuff you can do with them probably exceeds anything you might reasonably want to measure/control on typical home workshop machinery, and for those projects that require more, there’s always Mach 3 or starting from scratch.

 

But in general, I think the idea of a mech. eng. project bootloader sounds fine. Certainly saves a bit of work!

[Les Jones 1Participant @lesjones1]

Thanks to everyone for their comments.

I have concluded that most people that wanted a tachometer have already found a solution either by designing their own or adapting a cycle computer or counter.  As a result I do not think it a worthwhile project for the MEW magazine.

Les.

[Steve GarnettParticipant @stevegarnett62550]

I wouldn’t give up that easily, Les, because personally I wouldn’t judge the worth or otherwise of a project just based on a few responses in a forum – with the best will in the world, that’s rather a limited poll size! And anyway, if one can be made really cheaply, and with useful input options, then even people that have them already may suddenly find that they could use a few more, perhaps. And cheapness may well be the key here – the PIC you used only costs about £1, and although I haven’t priced up the rest of it yet, I really can’t see it being that expensive, even with a custom-made PCB. And on stripboard, even a bit cheaper perhaps, although that wouldn’t be my preferred option for making more than one of them, I must say.

 

For myself – well, the PIC programmer has arrived, and when I get a bit of free time (not so easy at present), I will investigate the whole thing a bit further.

 

So thanks for everything you’ve done so far Les.

[haroldParticipant @harold]

Hi Les

As one of the lurking majority I’m posting to echo Steve’s comments. I am sure there would be a lot of interest in your design amongst the unwebbed readership. Offering a stripboard layout option would widen the appeal to those without pcb-making kit and a list of Farnell/Rapid/RS order codes would simplify ordering enormously.

 

Please don’t bin the idea. It’s too good.

 

Regards,

H

[Les Jones 1Participant @lesjones1]

Hi Harold,

               I could probably draw a stripboad layout but it would not be as compact as a PCB. The real problem is it is not worth someone making a PIC programmer just for one project. What is needed is a magazine or company to be prepared to provide the PCB and programmed PIC (Or a kit). I do not think this magazine is set up to provide this service. I could program PIC’s for people but buying the components in small quantities is expensive. I used to buy components from Farnell but they have just increased their minimum order value to qualify for free postage from £20.00 to £45.00 so I will not be ordering from them as much. Anyone is welcome to the design and I would have no objection to someone producing kits as a sideline.

Les.

[Lawrie Alush-JaggsParticipant @lawriealush-jaggs50843]

What a thread eh?  It goes on forever.  Well not quite as long as the epoxy granite thread on CNCzone but a decent length anyway.

 

I decided to have a quick look with Google after reading Les Jones 1’s latest posting and found this.

http://translate.google.com/translate?js=y&prev=_t&hl=en&ie=UTF-8&layout=1&eotf=1&u=http%3A%2F%2Fwww.xbot.es%2Fmicroplans%2Ftacometro.htm&sl=es&tl=en

 

This is translated from Spanish to English.  This is the original site. http://www.xbot.es/microplans/tacometro.htm

 

As you can see, this is purspose designed for a lathe and has some interesting features including interfacing with a PC.

 

[Hugh GilhespieParticipant @hughgilhespie56163]

Hi Les,

 

I am another seconder (thirder?) of wanting to keep this project alive. If we could get a few other interested tach needers, I would be happy to order all the bits needed for however many tachos and then send them out at cost to the others if you would be happy to program a few PICs? A stripboard layout would also help as my experience of making PCBs is that by the time you have bought all the chemicals etc. it end up a bit dear for a single board.

 

Anyway, over to you!

 

Regards

 

Hugh

[Steve GarnettParticipant @stevegarnett62550]

Yes, the Spanish page is certainly interesting. 

The only real snag (and it is a very real snag) is the line that says:

 

“The program for the PIC is not available, anyone interested in construction can send the CIP schedule at an affordable price.”

 

Which seems to indicate that if you send him a PIC, he’ll program it for you and send it back – at what he determines to be an affordable price. Or, you write your own code. And on a technical note, I’m not convinced that using a 741-type device (rather obsolete) in the tacho preconditioning circuit is exactly a brilliant idea, even though it’s been used only as a threshold detector, but that doesn’t really matter very much.

 

But as you say, the features list is certainly interesting – well, a bit of it is. I like the idea of an approximate metres or feet per minute indication, but I’m not quite so sure what I’d ever use the rest of the facilities for. And if somebody could please enlighten me as to what the purpose/value of being able to send the speed data back to a PC is, I’d be quite intrigued.

Edited By Steve Garnett on 31/01/2010 13:51:11

[John RuddParticipant @johnrudd16576]

I too checked the website…Couldn’t see any pic code so gave up..

 

Personally I’d have used an LM393 or similar comparator for the pre conditioning cct…As for the remeote read on a pc…Superfluous imho…

[Sub MandrelParticipant @submandrel]

First: I’d like to see Les get his article published and then we can stand well back and see what happens!

 

Most micros these days have bult in comparators that will do the job as well as or better than an LM393, though a sensor (opto or hall effect) with built in conditioning is the most reliable solution.

 

Many micros have serial ports that can easily speak to a PC, just add a MAX232 or similar to get the right levels, or even a USB-RS232 convertor to allow USB connection.

 

The problem with a PC link is getting the PC to do something with the data. Can I plug BBC Basic for Windows by from Richard Russell – the best £30 i ever spent on computing related stuff. If you used to play with home computers years ago, then this will allow you to do all your old tricks and use windows interfaces without the pain and anguish of Visual BASIC or similar.

 

Neil

[John RuddParticipant @johnrudd16576]

Posted by Neil on 31/01/2010 21:01:11:

 

 

Most micros these days have bult in comparators that will do the job as well as or better than an LM393,

Neil

 

Sure, the 16F628A has a couple of compartors built in…but iunless you know how to use/disable them they are not much value to those that wield a turning tool…

 

Guys…Lets not get hung up on the why/wherefore etc of this project, just accept that someone has come forward with a worthwhile project….If electronics isnt your bag then dont criticise the OP for his offering..

 

For those that have experience, let’s not criticise Les for his efforts to make life simple…

 

Edited By John Rudd on 31/01/2010 22:46:55

[RoyParticipant @roy91521]

John,

 

I echo your thoughts, Thanks Les for the offer of a low cost, easy to assemble and functional Tacho.

 

For those of us into Electronics Design, we can either accept the project as is or if you feel you can improve then do so.

 

As for choice of micro, using the onboard comparitor is not difficult but will add overhead in code and time to execute instructions.

 

 

[Gerald WeareParticipant @geraldweare85524]

As one who has built a number of very successful tachos using the 16F628, the comparators are neither here nor there.  You need to have a digital filter in any case, so the hysteresis of the regular Schmitt trigger digital ports is fine.  In any event, I don’t believe the comparators have any hysteresis?  Been a while since I looked at them. Plus you probably need pull-ups for the comparators.  My tachos work just as well on a 16F84 as the ‘628.  All you need of the PIC is one input and enough output bits for the display.  What does seem to help a lot is to have an IR detector (or gear tooth sensor if that’s what you like) with a built-in Schmitt trigger.  Saves a chip.  I use nothing but the PIC standard I/O port and the sensor.

 

BTW, I can recommend the use of the FED (Forest Electronic Development)  PIC C compiler package.  It is full featured, including assembler, libraries for all the PIC subsystems, for driving LED and LCD displays and other devices, development environment, simulator, etc..  Very reasonably priced too, and good support.  Of course Microchip also has their free MPLAB environment which is fine, but last time I looked they were charging for the C compiler.

 

They also have an AVR product, but I have never used that chip.

[Steve GarnettParticipant @stevegarnett62550]

Posted by John Rudd on 31/01/2010 22:45:00:

 

For those that have experience, let’s not criticise Les for his efforts to make life simple…

John, I wasn’t aware that anybody was criticising Les at all – exactly the opposite, in fact.

[Les Jones 1Participant @lesjones1]

Hi all,

This reply is to Hugh Gilhespie’s  of Jan 31st

I think your offer is very kind as it would bring down the cost.

I did a costing of parts from Farnell and Rapid. Farnell worked out to £8.23 and Rapid £5.87 including VAT This is on the assumption that resistors could be bought in units of one. Farnell minimum is 50, Rapids is 100. For someone buying the parts for one unit (From Rapid ) they would have to buy 680 resistors more than required which is

£3.40 + VAT which is  £4.00 They would then have to Pay postage which is £4.95

(On orders under £30.00) The total cost for one would come to £14.82 which would make it too expensive. This does not cover the cost of stripboard  and bits of wire or a PCB A problem you would have is having some idea of the numbers involved. If the numbers were high then a manufactured PCB would be the best solution. I have tried to do a PCB layout with “Eagle” software but I have not yet mastered it properly. I would be happy to program the PIC’s If you want a copy of the costing spreadsheet send me a message on this forums messaging system with an email address so I can send you a copy. (I would not advise putting a email address in this part of the forum as it may result is you receiving spam.)

Les.

 

[Les Jones 1Participant @lesjones1]

Hi Neil,

           There is no reason why an internal analogue comparator should not be used.

It would require a PIC with three more I/O pins as I can see no way in firmware of internally routing the output of the comparator to the CCP input so it would have to be hard wired between the comparator output and the CCP input. I do not think this would improve the performance in any way but it may save some space. Removing an 8 pin IC but requiring going from an 18 pin PIC to a 28 pin one. I am making the source code available as well as the hex code so anyone can modify it to suit their requirements. (Or decide after looking at the code they do not like the design and go to another design.)

Les.

[Sub MandrelParticipant @submandrel]

Thanks Les,

 

Glad you took my post in the spirit intended. I’m not familiar with PICs and their architecture, but I can apprecaite the pin issue – driving displays eats output pins!

 

Why not look at hall effect switches? They contain a schmitt trigger and all conditionng circuitry, all you need is a +3 to +30v supply, and a pull up to Vdd on the output:

 

http://uk.farnell.com/allegro-microsystems/a1104eua-t/hall-effect-switch-sip-3/dp/1521705

 

I find one of these 100% reliable.

 

Neil

[Les Jones 1Participant @lesjones1]

Hi Neil,

             If you look back through this thread you will see that I suggested using hall effect gear sensor in the situation where sensing teeth on an existing gear on the machine is the simplest option. These have the magnet built in and automatic threshold adjustment so that the gear to sensor distance is not critical. There are two types of these. The two wire type requires a sense resistor in series with it (About 100 ohm is suitable.) The three wire type which would be ideal as it could be connected directly to the PIC carries a one off import charge of  £15.95 which makes them too expensive to buy in one off’s.

The part number of the two wire type is ATS642LSHTN (Farnell code 1521698) I used a device similar to the one you suggest with an external small magnet on my mill and that works well. For situations were there is no existing gear to sense then a reflective optical sensor is easy to install fitting it close to a rotating part with a reflective marker stuck to it. Both the two wire gear sensor and the optical sensor require an adjustable threshold. The design is aimed at non electronics people so I wanted it to be simple to use with either sensor. (If the unit was limited to using the exact hall device I suggest then the threshold could be set to a calculated value with fixes resistors.)

Les.

 

[Sub MandrelParticipant @submandrel]

Fair point on the gear sensors, but I wanted something that made one pulse per rotation, as I was timing one pules, rather than counting pulses in a tiome period. Same view, but looking down the other end of the telescope!

 

To be honest, I think half the reward of designing something like this is trying the various options – if they weren’t valid options then we wouldn’t have so many parts to choose from.

 

I think my bias for a magnetic sensor has more to do with the elevated oil and dirt levels around my lathe – I’m not sure a reflective marker would remain refelctive too long with me around

 

When you do your article you could consider giving examples of different sensor types and their pros and cons.

 

Neil

[haroldParticipant @harold]

Just noticed Les has updated his website with Tach Rev3a. On behalf of those like me, who are enthusiastic but ignorant: Thanks!

H

[Les Jones 1Participant @lesjones1]

Hi Harold,

                   Most of the changes on the website are as a result of things I find that need to be done before the information is in a form that can be submitted for publication. It is also a result of me trying to fully understand how to manipulate files on the website. At first I just used Microsoft Frontpage Publish function to create the website without really understanding what it was doing. The 3a version  software is a result of me noticing I had not put leading zero blanking on it. This has now been added. Also in the parts list the Kingbright display has been removed as the pin layout is slightly different to the Avego display I had used originally.

Les.

[Author]

Posts