Switching Led

[willeng]

That makes NO sense.

That doesn't surprise me.
What I was referring to is basically how can a Capacitor rated at 10V survive if it has 120V applied to it?


I didn't know about across the line & line bypass applications using different rated caps, thanks for the explanation.

I have a selection of Ceramic Caps.

What distance do you think I should mount the control circuit when finished away from the Universal Motors?
There is a lot of sparking at the brushes which I am not sure if or if not it will interfere with the control circuit if to close to them?

Just sorting things out now & i'll give it a go when finished.

Thanks again

Cheers

 

[KeepItSimpleStupid]

I'm guessing that C1 sees about 3 Volts DC. Do your simulation and put the probes across C1.

Anyway, you probably have both RFI and EMI to deal with, so put it in a metal grounded box. The RFI would come front he brushes and the EMI from the closeness to the motors.

You could put it in two boxes. The zero cross detector etc. near the motors and the "computer" elsewhere.

Aside:
RFI you eliminate by shielding. EMI you eliminate by twisting. This is why you see shielded twisted pair cables. The shield gets grounded at ONE END only which eliminates ground loops. If wires are twisted and exposed to a magnetic field and the set is differential, then whatever is induced in the first set is oppositely induced in the other wire and they cancel "in theory".

 

[willeng]

I'm guessing that C1 sees about 3 Volts DC. Do your simulation and put the probes across C1.

Yes, that's what I was trying to ask before when I mentioned about the capacitor having a higher voltage applied to it but as long as the voltage across it is less than 10V rating it's ok.

I just couldn't explain or understand it properly & said the differential voltage.

I normally use shielded twisted pairs etc wherever I think they are needed, I have found a suitable metal enclosure so everything is looking good.

I have to apologise for not adding LIKES to a lot of your posts, only because I didn't think about it, I am a novice also with the internet, forums, computers etc & don't use it all that much except for this Forum & that hasn't been much until now.

I understand about the Zero Cross Voltage Detection etc but some things just don't seem right to me.

In the Application note you posted on the Vacuum Motors it says that the Inductive load is NON LINEAR, I take that as meaning that the Current Lag may also vary hence the question enquiring about the variables that alter the Current Lag.

To me it appears that I am timing a pulse to the wrong thing, not for the Lamp being a resistive load but for the Motors & an Inductive load & a variable Inductive load at that.

I have noticed that the period of the Zero Cross pulse & the period of the gate pulse has an overall influence on the total range of the power output from the circuit, the wider the pulses are the circuit runs into trouble reaching the maximum output.

So instead of timing from the Zero Voltage Crossing & then applying a longer period pulse or pulse train etc to fire the triac after zero current crossing basically guessing where Zero Current is wouldn't it be wiser to just detect Zero Current & fire the triac with a short period accurate pulse & make achieving maximum output from the circuit a lot easier.
It would appear that then I am timing for the correct thing & not making a stab in the dark as to when the triac should be fired if you get what I mean.

I just seem to have this thing about timing for one thing & guesstimating for another that's all.
With engines timing is everything & also electronics as I am finding out now.

I have been playing around with a circuit that does detect zero current & it adjusts accordingly no matter how variable the Current Lag is, I have to improve it a little but it seems to be working nicely & is very accurate.

What do you think?

Cheers

 

[KeepItSimpleStupid]

So instead of timing from the Zero Voltage Crossing & then applying a longer period pulse or pulse train etc to fire the triac after zero current crossing basically guessing where Zero Current is wouldn't it be wiser to just detect Zero Current & fire the triac with a short period accurate pulse & make achieving maximum output from the circuit a lot easier.
It would appear that then I am timing for the correct thing & not making a stab in the dark as to when the triac should be fired if you get what I mean.

Not as I understand it. Way back when, I remember circuits that would use a transformer and actually generate a nice pulse at the zero cross. This was for wall clock timing, since line frequency triggering is has better average frequency.

Without that exact zero cross, your going to have to "make amends".

So, what the app note said is that you can expect up to 7 degrees difference in I=0 and V=0 with a universal motor, both of which will cause the triac to turn off. So, I suggested periodic pulses and the app note suggested to make essentially 100%, have a minimum trigger, It might also say that 98 to 100- is an invalid command IF 98% is the full half period-400 uS. So, that needs to be checked.

Because of the skew of the zero cross, your turn on pulse could cross at the other side.

This is what you have to balance. Don;t use continuous triggering because of power dissipation. Trigger within th the 7 degree mark 100% so if the a zero cross of current occurs within those 7 degrees, it's covered.

That firing window is very dependent on the actual frequency, the accuracy/symmetry of the firing pulse and the "real window". Your real window is shifted. When you zero cross pulse occurs, you can either be before the zero cross or after. If it's after, then you have to subtract that elapsed time. If it's before, you have to wait a bit before you start counting.

60 Hz or 50 Hz isn't exact. The power company can move it slightly. Maybe +- 0.5 Hz or so. I'd have to look it up.

 

[willeng]

Ok, I see what your saying.

I just spent a couple of hours searching for some info on the Zero Current Detecting for inductive loads & to my surprise I actually found something that may be useful?

See what you think, PDF Attached?

Cheers

 

[KeepItSimpleStupid]

I agree, it does look useful because it handles any phase angle differences.

It has a good zero cross detector and it deals with frequency issues

Three issues were not dealt with in that Application note:
1) V vs. V^2
2) Line voltage compensation
3) Current limit

We don't quite have a handle on #2 and I'm unsure how to implement #3. #3 probably isn't required.

Commercial devices usually split the triac system (triac/currrent limit) and might use isolated 4-20 mA for 0 to 100%. Some complex systems actually regulate power.

 

[willeng]

I'll study things a little & try to get an understanding of a few things in that application note.

By "Current Limit" to do mean inrush current on start up?

I am just spending a bit of time getting together everything I need as far as the MPX2010DP pressure sensor circuit goes, I will build the circuit this evening & calibrate it.

I have a few things to sort out properly & will post the results when finished, it will take a bit of time as I have to buy a few things to make a good job of it & for safety reasons.

I'll keep going & get back to you.

Cheers

 

[willeng]

Just a quick update on things,

I built the Pressure Sensor & Amplifier circuit & then set about calibrating this correctly which turned out to be a long drawn out process.
I wanted to calibrate the pressure sensor amplifier circuit to read 0.1V for every inch of water so at 40" of water it would output 4V to the Arduino.
I found that the Vacupulse unit I purchased to calibrate things had a very small fluctuation in the vacuum pressure, I'm talking about maybe .010" to .015" intermittent fluctuation on the Manometer of no particular frequency.
This made calibration virtually impossible as the mV range on the DMM wouldn't keep stable at all.

I tried larger vacuum lines hoping they would be large enough to act as a Plenum to remove any fluctuation etc.
I tried quite a few things & then decided to make a larger plenum to smooth out the fluctuation problem entirely.

It worked perfectly & is as stable as a rock & great for calibration purposes, in saying that the time now it takes to get precise measures on the manometer is very frustrating & it takes time for the pressure to stabilize for each step due to the large Plenum with many very fine adjustments via a needle valve to get it precise on the manometer.

At the end of the day it would appear to be calibrated near perfectly, calibration was checked & rechecked at every 5" of water & it is very good.

Do you think I should use a Low Pass filter on the output to the Arduino of this circuit, I did have a 4.7k resistor & a 10uF cap as a filter on the output but removed it?

I then made up a multiple Set Point Circuit basically it has 8 outputs of 0.5V to 4V (5" to 40"water) in 0.5V steps (5" water steps) using a multiple position switch.

Although not wired in properly yet I have attached a picture of the circuits mounted in a steel enclosure, I am going to mount the Arduino Board in this same enclosure & keep the Zero Cross & the Triac circuits separate in another enclosure.

I am progressing slowly & am learning a few things along the way.

If you pick anything that will cause issues please say so & I'll correct them?

Cheers

 

[MrAl]

Hi,

One thing i noticed right away about the circuit board is there are not enough pots on the board. 10 trim pots are not enough, to get 1 percent control you need at least 100 pots

Seriously, i wonder what all those pots are doing and why they are not mounted at the edge of the board so they can be adjusted easily.

 

[KeepItSimpleStupid]

Yep, the orientation of the pots looks dumb or you have 20/20 hindsight or whatever perfect eyesight is in metric form.
So, why not use course and fine adjustments. e.g 10K+500 in series or whatever.

I know the feeling. I want to simulate a 50K at 25 C thermistor, so I made a box with 6 pots: 50K, 20K, 20K, 10K, 500 and 100 ohms all with turns counting dials. I'm planning to add a 100K pot. I'm not sure if the 7x wiper resistance will be a problem.

 

[MrAl]

Hi,

Why do you want to simulate a 50k thermistor?

 

[KeepItSimpleStupid]

Mr Al:

OK, now back to a rather unfinished project. The micro/convection oven, First, it has to be fixed. A triac and opto+series resistor fried and I somewhat forget the circumstances, but the magnetron popped and wasn't replaced in time. The repair forum where someone posted a request for help on my very same micro/convection oven is apparently no more. I think I tried putting a small bulb for the magnetron and the other stuff popped.

I guess I'll have to take some pics. In any event that was a early Kenmore/sanyo convection/micro built around the mid 80's. Searspartsdirect did have a parts list and a stuffing diagram at one time that I nabbed, I finally found service manual, so I have a parts list which is not in the SM; the SM which does not have a schematic. It;s hard to figure why board level parts were available at one time and the SM doesn't contain either the parts list or the board schematic. It's still unfixed. Convection works. Could be unfixable if the processor port is damaged. The IC seems to be available if I could speak Chinese.

In any event, the oven is huge and it had three design faults. Sanyo fixed one and attempted a fix for the other which I corrected. It supported the fan with one bearing. It's now supported with two. They changed the shaft to be non heat conductive.The oven was like $950 in the 80's. Sanyo fixed the rack supports. Suffice to say, I really like it.

The third fault: What's peculiar, is how the temperature probe is designed. It uses a modified 1/4" phone plug with a large insulated area for the independent actuator contact that rests on and the plug is not compatible with normal phone plugs or high temperature ones. The high temp ones would be difficult to mount and don't include the switch. The probe operates from 105 F to about 180 F or so. 50K NTC at 25C.

I figured out what curve it is. It's a Y with 50K at room temp. Luck would have it, I managed to get a replacement probe with the IDENTICAL characteristics . The OEM probe has an intermittent contact at the probe. Both the new and OEM probe were characterized from about freezing to above boiling using a potato and (water or oil) and a stove. Try getting this info from the guys making the replacements. GE tried, but said take the data with a grain of salt which I did. It was wrong.

I managed to also get a 1/4 phone jack with an isolated contact because the original jack is damaged upon recent inspection. It happens to have a Nylon actuator which isn't a good choice because heat up to 450 F escapes through the 1/4 hole and passes the Nylon actuator. This isn't good long term.. So, a few possible fixes now presented itself. For all fixes, the thermal conduction to the jack phenolic will be reduced with Titanium washers.

Th easiest fix, is to use a piece of Macor (a machineable ceramic) to plug the hole when the probe isn't in use. The second and third options are dying fast because my access to a lathe could disappear at any time. A friend is in the process of dissolving his business. The lathe was used primarily to keep the old machinery running and to make jigs.

Other options are to replace the Nylon with Macor and either glue or try to use a plastic thread forming screw. All three should work. The latter two have convenience going for it. I think I have a suitable screw.

So, it looks like I have a fix.

So, I should have a working probe and a convection/microwave that's fixable. I have a new magnetron and HV capacitor.

The ORIGINAL intent was two fold:
1) Make a thermistor type changer with open thermistor detect.
2) and to TEMPORARILY change the range to 70 degrees less. e.g. 80F + 25 is 105F and 105+80 = 185F so I can include at least room temperature and yeast proofing temp which is like 95 to 100 F although these **broken link removed** guys think differently than https://en.wikipedia.org/wiki/Proofing_(baking_technique) A sloppy 105 F is too hot.

So, why can't I warm something from the fridge or freezer to room temperature?

The thermister curve is nuts.

Some numbers:

F C K-ohms

50 10 99.72
68 20 72.84
86 30 53.01
95 35 42.92
104 40 32.55
113 45 26.20
122 50 21.75
131 55 17.94
140 60 15.38
158 70 11.39
176 80 8.57
194 90 6.84
203 95 5.98
212 100 4.88

I need the simulation box to figure out what the 105 F to whatever in steps of 5 degrees is in terms of voltage across the thermister. I haven't reverse engineered the thermister input circuit yet. It is a thermister to ground, bit that's about all I know, I think the open circuit voltage is 5V.

Ultimately, I want to replace the thermister with an OPTO-FET and a totally different linearizion to accommodate room temp and the yeast proofing temperature. I envision something like push a button and the low temperature mode is entered. So, if I want 75 deg, I'll set the temperature to 75+80.
When the probe is pulled out, the OEM mode is reset. Readings would be off too.

Initially, I was thinking of a thermistor type changer with two ranges and open thermistor detect, but I don't really have to do that.

I think I got th MAX1452 and the corresponding development kit **broken link removed** but I don't think it's going to work out. This part the TI PGA309 https://www.ti.com/product/pga309 might be a much better choice. Maxim agrees that their documentation needs work, but it isn't going to happen.

So, I don't care if every temperature below room temperature is room temperature either. Right now, every temp below 105 F is 105 F.
In order to actively do anything, I need the voltage across the thermistor at 5 degree increments which is the resolution of the setpoint and the measured value. I will definitely include an indicator even if it's on the back of the microwave by the push button when the low temp mode is activated.

It gets messy especially with open thermistor detect. I don't know if there would be any other glitches. I can't use a 100 deg F offset.

The machining is conceptually easy. I need to machine a shoulder, but it's a small part and I've never machined Macor before. I found an epoxy that would work, but it's also way to expensive, thus the idea about the screws I already have. You can tap Macor, but this screw is tiny. I've procured the Macor for both options.

 

[willeng]

Hi,

One thing i noticed right away about the circuit board is there are not enough pots on the board. 10 trim pots are not enough, to get 1 percent control you need at least 100 pots

Seriously, i wonder what all those pots are doing and why they are not mounted at the edge of the board so they can be adjusted easily.

Hi MrAl,

The Eight Pots in a row are to trim the eight voltage dividers to give 0.5V to 4V output in 0.5V increments to be used with a multi position switch for 8 accurate Set Points.
Yes, they could have been mounted on the edge of the board but for this "prototype board" I decided to mount them where they are to keep all unnecessary wiring away from the Arduino, that's the only reason. They are easily adjusted.
I could have used a multi turn pot to get my set points but I have trouble now with balance issues etc making some things very difficult to achieve, so I opted for eight set points only which is all I require & at the turn of the multi position switch it makes life easier for me now with one click at a time.
The other two pots are to trim the INA126P Gain & Offset.

Could you give me some advise for a way to do it without using so many pots, resistor values are not accurate enough to get good results & they require trimming?
I need eight accurate set points, so I just used voltage dividers & trim them with the pots?.

Hi Kiss,

Not really sure why I need Course & Fine adjustments, I just need eight accurate set points as mentioned each at 0.5V increments?
Do I need a low pass filter out of the INA126P to the Arduino input.

Giving way to aesthetics I thought it better to route wiring away from the Arduino rather than have something look good but run wires all over the place creating problems, my bad if it doesn't make sense or look good.

Like I've mentioned before I am only a beginner with electronics & not ashamed to admit it?

Cheers

 

[KeepItSimpleStupid]

Does accurate mean "spot on" or just repeatable? I think you did OK. You just need a crooked screwdriver to adjust the pots. No big deal at all.

The only thing I would not do, is use the power supply for the divider reference unless whatever is reading it is "Ratiometric" and not absolute.
Reference IC's are more stable. I "think" your pressure sensor has a ratio metric output.

So, if a processor reads the vacuum sensor, the output is relative to the supply voltage. D/A references can sometimes be changed so the power supply is the reference.

The aesthetics look great. The ergonomics could use a little help.

==

Think of it this way, I think your pressure sensor's output is relative to the supply voltage, so your setpoints also need to be relative to the supply voltage.

So, when the supply is 5.0000 V, you may want 0.5000 and 1.000 volts, but if the supply drops to 4 V for whatever reason, you don't want those same voltages. So, for the 0.5V case, it's 1/10 of 5V. So with a 4V supply, you want 0.4000 volts or 1/10 of the supply. Any noise in the supply also ends up to be ratioed as well. The sensor would see the same changes and the 0.4 and 0.5V would be equivalent. V supply is a function of time. A 5.000 reference is not. By making the reference and the sensor affected by power supply variations, the variations dissapear.

It isn't the same as having a sensor that has a 0-5 V output for say 0-50 engineering units.

 

[willeng]

Does accurate mean "spot on" or just repeatable? I think you did OK. You just need a crooked screwdriver to adjust the pots. No big deal at all.

The only thing I would not do, is use the power supply for the divider reference unless whatever is reading it is "Ratiometric" and not absolute.
Reference IC's are more stable. I "think" your pressure sensor has a ratio metric output.

So, if a processor reads the vacuum sensor, the output is relative to the supply voltage. D/A references can sometimes be changed so the power supply is the reference.

The aesthetics look great. The ergonomics could use a little help.

Haha, no crooked screw driver required, I have very small watch makers screw drivers that fit perfectly, I looked at that before mounting things.

Spot on & repeatable is the go.

Yes, I was also concerned about variations in the power supply & it's effect on the divider reference voltages, pressure sensor supply & the INA126 so I loaded the power supply output with more than double the load it will encounter with all the circuits combined & checked to see if any variation existed but I found no variation at all.
The power supply I am using is from a computer & is rated far above what is required but I have several of them to use up.

I see no variation in the output voltages to the pressure sensor or anywhere where else even though the mains voltage does vary.

Yeah your probably right I could have done better but by the time I fit everything into the enclosure including an LCD, Led's, Switches, Circuits etc I was more concerned about the wiring & where it is to be routed rather than anything else.

Yes, reference IC's would be better, but I am stuck with what I have available at hand which is a lot of recycled components, like on the circuit shown, besides the pressure sensor, resistors & some of the capacitors which are new.

I think the biggest thing to remember is that no matter how accurate I make things, the Flow Variations of certain test pieces are sometimes not stable at all & sometimes erratic turbulent flow exists before modifications are made to the test pieces to smooth out the flow.

Sometimes the test pressure can increase & decrease rapidly along with total flow through the test pieces, this is of more concern as the PID will try to correct for the variations which are unavoidable, in fact you try to get all the test pieces unstable at first so you can see what the maximum potential flow can be once configured correctly.

Airflow through the test bench can be a violent affair as it is through engines.

I'll have a manual override switch so I can use the test bench manually when unstable flow exists, then when sorted I will switch over to the PID circuit to get final results or otherwise the PID will create havoc trying to correct for unstable flow conditions which can be very unpredictable to say the least.

Anyway I'll keep going at it & see what I come up with.

Cheers

 

[MrAl]

Hi,

Well in order to answer the question posed here about creating accurate voltage dividers, we first have to know what the tolerance really has to be on each voltage. You really have to think about this because if you shoot for too good of a tolerance it may be just an overkill and then it is difficult to achieve, while a more practical tolerance will be easier to implement.
For example, if you look at your circuit and find that the readings (or whatever) all come out close with 2 percent tolerance, then dont quote a 1 percent tolerance. Likewise, if 1 percent is good enough dont quote 0.1 percent. For most applications 1 percent is good enough and 0.5 percent is way good enough.

There are more precision resistor values too, like 1 percent and 0.1 percent. That could mean just ONE trimpot for any number of voltages.

On the other hand since you already have this built maybe it's good enough as it is, unless you want to reduce size or make more of them.

 

[willeng]

MrAl, thanks for the reply,

I see what your saying with the tolerance, I did spend time setting exact voltage references & have for example 0.500V, 1.000V, 1.500V, 2.000V etc, this is with my DMM of course & who knows how accurate that is anyway. I was thinking that things have to be very accurate but now realise differently after your post.
I did the same with the output from the Pressure Sensor amplifier as well and I see that I spent far to much time trying to get exact readings but they do appear to be on the money.

Yes, precision resistors would be good but unfortunately I only have 5% resistors at hand at the moment.

I will build the whole thing properly as this is only a prototype to see if all goes well to start with, it's a big learning curve for me & thanks to the input from others I will learn how to do things better or in my case, correctly.

Cheers

 

[KeepItSimpleStupid]

For giggles, post a link to the datasheet for the pressure transducer. I'm interested in the transfer function. It probably contains a Vcc term.

An example product, not available, of course, is say 10 isolated 1K 0.1% resister array. **broken link removed**

 

[MrAl]

MrAl, thanks for the reply,

I see what your saying with the tolerance, I did spend time setting exact voltage references & have for example 0.500V, 1.000V, 1.500V, 2.000V etc, this is with my DMM of course & who knows how accurate that is anyway. I was thinking that things have to be very accurate but now realise differently after your post.
I did the same with the output from the Pressure Sensor amplifier as well and I see that I spent far to much time trying to get exact readings but they do appear to be on the money.

Yes, precision resistors would be good but unfortunately I only have 5% resistors at hand at the moment.

I will build the whole thing properly as this is only a prototype to see if all goes well to start with, it's a big learning curve for me & thanks to the input from others I will learn how to do things better or in my case, correctly.

Cheers

If you have 5% carbon resistors one trick we used to use is to start with a lower value (not too much lower) and cut a "V" shape in the mid section with a small "V" shaped file. As you measure some quantity like the voltage, gently file a little deeper into the body of the resistor. As you dig deeper, the resistance goes up and that is how you adjust the voltage. Finish up with a coat of electrical clear coating type paint.
You can adjust the value this way but of course the temperature tracking wont be as good as a 1 percent or 0.1 percent resistor, and you might want to try it with carbon film type too first to see if this works with them too.
Not the easiest to do but it's cheap

 

[willeng]

MrAl,
Yes, that's one way to do it but pots seem to be the go at the moment, with the health issues I have now I couldn't imagine myself attempting such delicate work.
Besides brain damage among other things I am a bit unbalanced with things seemingly in constant motion & very fine work I am not good at now, it's a real pain in the butt, hence the need for automated control of my test bench. Even setting a multi turn pot to exact values is an arduous task.

Kiss,

Here's the Datasheet for the pressure sensor.
Interested to hear your thought's on it?

Cheers