The Lowspec spectrometer

[Thalestris24]

1 hour ago, andrew s said:

If your using bass and you save the calibrated image the calibration will come back if you load the saved image. If you make it No 1 in the sequence then you can apply it to new spectra. 

Regards Andrew 

Hi Andrew

Yeah, BASS. I did some sort of a save but it doesn't appear to have saved properly. It wouldn't reload the saved file which was a .fit but I think only a fraction of the size it should have been. I obviously did something wrong! I would have thought there would have been a warning when I chose to remove it from the display. Oh well.

Louise

[Thalestris24]

Oh, wait a minute.... I just reloaded the image and the chart does actually come up with wavelength rather than pixels but there are no annotated calibration lines. Should I be doing something like saving it as a reference?

Louise

[ejp1684]

Go to Calibration on the top menu and select 'Show/Hide Calibration Points'. That should toggle the calibration lines on and off.

Eric.

 

[Thalestris24]

5 minutes ago, ejp1684 said:

Go to Calibration on the top menu and select 'Show/Hide Calibration Points'. That should toggle the calibration lines on and off.

Eric.

 

Ahhh….

[Merlin66]

Louise,

I take a reference exposure every time I take a target spectrum of a star....

It quickly becomes second nature.......

[Thalestris24]

On 28/12/2019 at 22:50, Merlin66 said:

Louise,

I take a reference exposure every time I take a target spectrum of a star....

It quickly becomes second nature.......

Yes, think you said that before. Does that mean you recalibrate from scratch every time - even if no settings have changed? Or does your grating position always change meaning you have to calibrate?

I've been busy designing a calibration starter lamp holder that will just simply fit via a hole in the lid. I just need to get a drill bit that will let me make a 16mm hole. Will try and get that tomorrow. Then I'll put the design into Fusion360 and print it out .

Cheers

Louise

[Merlin66]

Louise,

Yes, every time... thermal  effects, flexing in the housing etc etc etc can impact on the accuracy of measurements. It's not unusual to take a reference image before and after a long series of target exposures.

Fitting a reference lamp - check the alignment and orientation to the slit gap. Ideally the lamp should be on the optical axis as near the slit gap as possible.  On the Spectra-L200 I used a flip mirror to direct the reference lamp to the slit gap.

 

[Thalestris24]

5 minutes ago, Merlin66 said:

Louise,

Yes, every time... thermal  effects, flexing in the housing etc etc etc can impact on the accuracy of measurements. It's not unusual to take a reference image before and after a long series of target exposures.

Fitting a reference lamp - check the alignment and orientation to the slit gap. Ideally the lamp should be on the optical axis as near the slit gap as possible.  On the Spectra-L200 I used a flip mirror to direct the reference lamp to the slit gap.

 

Ok. I've designed it so that it will be just in front of the slit and with the lit part about opposite half way up the slit. Should be ok. Before, I had it temporarily fitted via a t2-1.25" adapter at the front. It should be a lot better coming down from the lid. I might make it with notches so it's in the same place each time it's used.

Cheers

Louise

[Merlin66]

Lousie,

Now that Paul's on board, you should ask him when he's going to incorporate a reference lamp in the design of the LowSpec

[Thalestris24]

9 hours ago, Merlin66 said:

Lousie,

Now that Paul's on board, you should ask him when he's going to incorporate a reference lamp in the design of the LowSpec

Yeah, but how do you standardise a calibration light source? Except by using an expensive, dedicated one? The Alpy one is a possibility. The S10 bulb I'm using seems ok but the starter bulbs from different manufacturers are all likely different and I haven't finalised the emission wavelengths of my S10 yet. There is also the issue of powering one like the starter bulb. I have experience of electronics so I'm comfortable wiring a starter bulb up to the mains and incorporating a suitable current-limiting resistor (soldering required). But others may not be so comfortable and the uk mains supply is potentially very dangerous.... Maybe it would be better to use a 12v inverter or a cold cathode psu though high voltages are still involved. Perhaps just using an external light source such as the cfl might be good enough for most (so long as they are available). Some thought required.

Louise

[Paul Gerlach]

I'm also struggling with designing a calibration unit for the LOWSPEC that is easy to build for someone like me who has no experience in building electronic circuits.
I've recently ordered a 12 Volt DC to 220 Volt AC converter board form our friends in China. Maybe this can bring me up to speed.

[Thalestris24]

6 minutes ago, Paul Gerlach said:

I'm also struggling with designing a calibration unit for the LOWSPEC that is easy to build for someone like me who has no experience in building electronic circuits.
I've recently ordered a 12 Volt DC to 220 Volt AC converter board form our friends in China. Maybe this can bring me up to speed.

Hi Paul

I was considering getting something similar but that's a 150W power supply for something that draws maybe a 1/3 of a Watt or less! The output of the PSU is still a dangerous voltage though using one of the lower voltage tappings might be ok. You still need a current limiting resistor. To be honest, I'm not entirely sure what, for our purposes, the working voltage of, for example, a starter bulb might be. 

Louise

[Paul Gerlach]

Louise,

Hmm you may be absolutely right. Shows that I really need some help in this department 😉
I believe the working voltage of the starter bulb is about 80-90 Volt.

[Thalestris24]

2 hours ago, Paul Gerlach said:

Louise,

Hmm you may be absolutely right. Shows that I really need some help in this department 😉
I believe the working voltage of the starter bulb is about 80-90 Volt.

You may be right about the working voltage but think it's actually higher. There is a formula which relates resistance, current, supply voltage and operating voltage:

R = (Vsupply - Vbulb)/I where I = current, of course.

In the UK I think we take our mains supply to be 240V rms.

I'm using my Philips S10 starter with a 27k resistor in series and the two connected directly to the mains. Seems to work!

If we assume the circuit draws about 2mA and rearranging the equation:

Vbulb = Vsupply - IR (IR = voltage across the resistor)

Vbulb = 240 - (.002 x 27000) = 240 - 54 = 186V

Were it 3mA instead:

Vbulb = 240 - (.003 x 27000) = 240 - 81 = 159V

That gives us a back of an envelope range to work with.

As the bulb and resistor are in series, the current through both is the same. It's a bit more complicated than that in reality because the bulb has a dynamic resistance. The nature of a starter bulb is such that initially it is open circuit but rapidly closes with heating to a short circuit - hence the absolute need for a current limiting resistor. Also, it's resistance and current flow vary dynamically. If there were just the resistor across the 240V mains it would pass:

I = V/R = 240/27000 = 8.9mA.

The heat dissipated is given by I^2 x R 

Just the resistor:

Edit: (0.0089)^2 x 27000 = over 2W!

Well, it's only a 0.25W resistor... If it were on its own it would probably catch fire rapidly ha ha.

At 2mA the dissipation in the resistor will be

(.002)^2 x 27000 = ~0.11W

And at 3mA it's 0.24W. The whole thing does get warm but not too hot to touch (I'll double check that when I can!)

By the way - thank you so much for the Lowspec design! Much appreciated

Louise

 

 

 

 

Edited December 30, 2019 by Thalestris24

[Thalestris24]

I just did a quick test. The glow was steady for about 5 mins. Thereafter it flashes on and (briefly) off roughly every second-ish. There may be a case for using a slightly higher value resistor - maybe 33k. Oh, I remembered my resistors are of the metal film type so should be able to dissipate up to 0.4W. Always best to be conservative, though . I don't know what the minimum current is which results in the bimetallic contact closing and switching the bulb off. If my one was running at a slightly lower current then maybe it would stay steady... I don't have any 33k resistors in my box of bits so I'll have to order some in and try them out. It doesn't actually matter that much that the bulb closes the circuit briefly - you can still get a spectrum ok. However, every time it closes that means more current will pass through the resistor which could get hot. Again, maybe not a problem in practice.

If there is a 33k resistor with 240V across it:

I = V/R = 240/33000 =  ~7.3mA and I^2 x R = 1.76W which is quite a lot of heat. 

Although my 27K resistor seems to work, there may be a need to increase its power rating, just in case. It all depends on how long the contacts may be closed for. On the other hand, if the current is low enough with a 33k resistor then the contacts may stay permanently open - great! There is only one way to find out how the starter bulb behaves so I'll try it with some other resistors. Well, as it's a soldering job, I'll just try it with a 33k resistor first. According to a Farnell datasheet a neon bulb current can vary from 0.3mA to 2.9mA, depending on the particular bulb so I think assuming 2 to 2.5mA probably puts us in the right ball park for the starter.  

Louise

[ejp1684]

 

With the bulb just in front of the slit the exposure time will be much  quicker; mine for the S10 are typically 5secs to 20secs depending on which end of the spectrum I'm using.

I'm not very good with electronics so bought this inverter from ebay which plugs into my battery box.

I used resistors in series with the S10 bulb (33k from memory which I think gave about 86v) and an inline fuse, and the S10 flickers slightly but remains on as long as power is connected. My 'hole in the lid' of Lowspec works, but not ideal. Would be great to have something like a flip mirror with the bulb fixed inside the case.

Is it possible to get an 'adjustable' inverter to provide approx. 86v from a 12v supply?

And as Louise said, many thanks for Lowspec, Paul.

Eric.

[Thalestris24]

So having come across the neon bulb data I thought 'I wonder if there's any data for glow starters?'. Don't know why I didn't look before - d'uh. Anyway found this for Osram starters. I also found they have a closure and a non-reclosure voltage. The closure voltage is the value that closes the contacts. The non-reclosure voltage is the threshold voltage below which the contacts won't close. Very useful!  The non-reclosure voltage for the S10 is given as 140V so ideally need a voltage below that - maybe 130V. (btw Farnell say the S10 is no longer manufactured - this is likely to be an ongoing problem for starter bulbs). Anyway, from the equation I mentioned above:

R = (Vsupply - Vbulb)/I 

R = (240-139)/.002 = 110/.002 = 50500

That looks like a 51k would be nearest value.

If the current were 3mA it would give 110/.003 = 36700

So worth getting some 39k, 47k, 51k and 56k resistors to try.

Btw, I'm thinking the bimetallic strip would likely never stay closed for any length of time, or fail in a closed mode, so a short circuit through the bulb probably won't ever be a problem.

Louise 

[Thalestris24]

41 minutes ago, ejp1684 said:

 

With the bulb just in front of the slit the exposure time will be much  quicker; mine for the S10 are typically 5secs to 20secs depending on which end of the spectrum I'm using.

I'm not very good with electronics so bought this inverter from ebay which plugs into my battery box.

I used resistors in series with the S10 bulb (33k from memory which I think gave about 86v) and an inline fuse, and the S10 flickers slightly but remains on as long as power is connected. My 'hole in the lid' of Lowspec works, but not ideal. Would be great to have something like a flip mirror with the bulb fixed inside the case.

Is it possible to get an 'adjustable' inverter to provide approx. 86v from a 12v supply?

And as Louise said, many thanks for Lowspec, Paul.

Eric.

I think from my calculations it may be better to aim for a bulb working voltage of over 100V but below the quoted non-reclosure voltage threshold of 140V. I'm only likely to ever use my setup indoors, so running directly from the mains isn't a problem for me. I have (briefly) looked but it seems hard to find an off-the-shelf low power inverter that will do the job. The one Paul quoted above would actually work ok - in a suitable enclosure. It has output tappings for 90V and 110V. You'd have to recalculate the resistor value to use at the lower supply voltage. It's a 150W (that's Chinese Watts...) so a bit overkill and has no short circuit protection so you'd have to put a fuse in the output. Still I think it could be used ok. There are still high voltages - up to 220V on the output. Maybe there are other inverters out there that would be better. I'll have a look on Ebay Something with a 110V output might be good. Your working voltage of ~86V may be fine, Eric. That would suggest the bulb current is higher than I'd assumed. I'll have another look at the calculations based on, say, a 90V working voltage and working from that .

Louise

Edited December 30, 2019 by Thalestris24

[Thalestris24]

Quick calc based on 33k and a 90V working voltage gives a current of 4.5mA which seems quite high (quite a bit higher than the neon bulb currents quoted by Farnell). Also gives a power dissipation of 0.68W so would need a 1W resistor and would get very hot. Is that what you used, Eric? Doesn't seem to agree with my use of a 27k resistor which is close in value to 33k. I'm not too keen on measuring mains voltages although my old digital  multimeter will do it ok - I just don't trust myself! It might make more sense if that voltage of 86V was across the resistor rather than the bulb. That would give a current of 2.6mA - in my ball park of 2 to 3mA. Of course, I could be quite wrong!

BTW, I noticed that the above inverter outputs are high frequency, 20kHz square wave. Not sure if that would affect the operation of the bulb.

Louise

[Thalestris24]

I found this little inverter on Ebay. I ordered one from an alternate listing which said it was located in UK but I'm not convinced!

Anyway, if I've read the Chinglish right, it says it can output AC 100V, DC 220V, AC 220V from various DC inputs 3v-14V. Not clear exactly what it will do but it's small and cheap! Would need a suitable enclosure. Soldering required. Hope it comes with some wiring instructions, lol. 

@ejp1684 I'll let you know if it's any good, Eric  

Edit: some info on Banggood:

Louise

Edited December 30, 2019 by Thalestris24

[DaveL59]

yikes, don't like the fact that there's no drillings for mounting posts in the board nor any heatsinks. I guess for the price it is dirt cheap, make sure its in a well insulated enclosure and cables secured with suitable strain relief. You might want to have some ventilation also to assist cooling if running for extended periods, guess you'll know once its up and running and you can carefully check its operating temps.

[Thalestris24]

4 minutes ago, DaveL59 said:

yikes, don't like the fact that there's no drillings for mounting posts in the board nor any heatsinks. I guess for the price it is dirt cheap, make sure its in a well insulated enclosure and cables secured with suitable strain relief. You might want to have some ventilation also to assist cooling if running for extended periods, guess you'll know once its up and running and you can carefully check its operating temps.

Hi Dave

Are you referring to the board I just posted about? It's only a low power board, about 4.5cm square. Will need to go in a small box and would probably benefit from a fuse/switch/LED indicator etc. It's only to power a starter/neon-like bulb so no power drain involved - it's just a way to get a higher voltage from a 12V, or less, input, and to power the bulb without a direct connection to the mains (as I have at the moment).

Louise

[DaveL59]

Hi Louise

yeah that was the one, at that low output hopefully it'll be fine then  Must admit I looked at the photos of its build "quality" and toes just curled. You'd expect a better finish in mass produced boards but then at that price... Having said that the rats-nest wiring in some items at much higher prices leaves me speechless too, surprising what can lurk beneath a nice classy fascia that you don't get to find out till something stops working.

[Thalestris24]

Just now, DaveL59 said:

Hi Louise

yeah that was the one, at that low output hopefully it'll be fine then  Must admit I looked at the photos of its build "quality" and toes just curled. You'd expect a better finish in mass produced boards but then at that price... Having said that the rats-nest wiring in some items at much higher prices leaves me speechless too, surprising what can lurk beneath a nice classy fascia that you don't get to find out till something stops working.

Yeah, it certainly looks very basic - hardly quality pcb or components. I think it's aimed at people that have, at least, some idea of what they're doing! It should work ok from a low power mains psu. Might even be ok with a 5V or 12V Li-ion battery. It only has to run long enough to take an exposure of the calibration bulb
Louise

[Paul Gerlach]

Well, the ALPY calibration unit obviously has a an inverter build into it as it uses 12 volt and a Relco.
In the book 'Spectroscopy for Amateur Astronomers' from Marc F.M. Trypsteen and Richard Walker there is in appendix D a circuit depicted that converts 12 Volt DC into 230 Volt AC.
So in order to avoid messing with a soldering iron the 'Chinese circuit' should do the same.