Measuring density with ultrasonics

[Tony Giaccone]

I'm interested in determining the density of a dough. I want to be able to continuously measure the density of the dough as it rises. My thought was to use an ultrasonic wave and measure the time it takes to reflect through the dough I would assume that as the dough ferments and the amount of CO2 in the dough increases, and the size of the dough ball increases that the density can be measured, but I'm not sure this will work. Has anyone got any advice or suggestions on the best way to accomplish this task?

 

[Dale]

That is an interesting idea. I assume that you are thinking of this in terms of a process control?

I think that the dough/air interfaces would scatter the ultrasound more than anything. So you would probably be able to make some measure of scattering that you could use in your process control even if it wasn't actually density.

 

[Tony Giaccone]

That is an interesting idea. I assume that you are thinking of this in terms of a process control?

I think that the dough/air interfaces would scatter the ultrasound more than anything. So you would probably be able to make some measure of scattering that you could use in your process control even if it wasn't actually density.

So my brother in law runs a pizza business, and I'm trying to identify at what point a ball of dough is best suited to being turned into a pizza. There are a lot of factors that influence when the dough is "ready" and density feels like it might be a useful measure. There's a honeycomb of C02 and dough, and also % of water and I bet ph as well. At this point I just want to get an idea of what we might easily do.

 

[Dale]

You should be able to get sensors directly designed for pH and moisture, and of course temperature.

Ultrasound may be a little cumbersome and expensive for measuring density. If your pans for raising the dough are standardized then you could just use a webcam to measure an outline and infer a volume from that.

 

[sophiecentaur]

You should be able to get sensors directly designed for pH and moisture, and of course temperature.

Ultrasound may be a little cumbersome and expensive for measuring density. If your pans for raising the dough are standardized then you could just use a webcam to measure an outline and infer a volume from that.

Ultrasound (or at least 'sound') is convenient to use, these days. Measuring the reflection coefficient in air[edit: not, not air. It would have to be more dense than that] at the surface would involve a head, placed against the surface of the dough. As long as the wavelength of the sound is significantly longer than the pitch of the foam, the dough could be treated as a bulk medium.
But your optical suggestion would have the advantage that it wouldn't require any development of hardware - just some software to examine an image of the dough ball. One simple way to measure volume could be to put a single sample of a batch of dough in a cylinder and use a light meter to measure how much the risen dough obscures a light source - to measure its volume. That would involve getting an accurate size (mass) of dough to start with though.

How about rising the dough balls on a translucent tray , placed over a light source and monitoring the light getting to an overhead sensor? Area at the beginning would be compared with area at the end, to give the increase in area - implying volume. That would rely on fairly consistent behaviour of the dough balls but wouldn't require them to stay spherical. Some calibration runs would be necessary to check for consistancey of measurement. Could be an easy experiment to set up (all equipment is readily available).
But the idea of measuring an Intrinsic quantity is attractive - hence the possible advantage of finding density.

 

[Tony Giaccone]

The dough balls are 8 to 12 to a pan and they are stacked in trays. My thought on the ultrasound was you could physically touch the dough all with the ultrasound head and use that to do a quick, red, yellow green assessment of the dough. The problem is that the ambient conditions in the "kitchen" vary greatly. Calculating the volume can be tricky as the dough balls touch each other as they expand. I can buy an ultrasound transducer at radio shack for about $30 which is designed to measure distance so it may not be useable in this way. My other though was to get a piezo electric "speaker" (I don't think the frequency goes high enough). But one on either end of the ball might be interesting. Measure the time for a sound wave to traverse the ball. I am not sure if there would be an interval long enough to measure with any accuracy.

 

[sophiecentaur]

I can see that an (ultra) sound method would be convenient as it would work on any shape of dough ball. A measurement of the reflection coefficient could tell you the density. You could measure the ratio in amplitude of the forward and reflected pulse. If you're trying to develop something like this then you would need a scope at least.
I did a bit of googling and found that most of the ultrasound applications thrown up are medical. I did find this link which would be worth reading through for some background.

 

[nasu]

So my brother in law runs a pizza business, and I'm trying to identify at what point a ball of dough is best suited to being turned into a pizza. There are a lot of factors that influence when the dough is "ready" and density feels like it might be a useful measure. There's a honeycomb of C02 and dough, and also % of water and I bet ph as well. At this point I just want to get an idea of what we might easily do.

Using ultrasound is hardly something that will count as "easily do" in this case.
I mean, to get something meaningful you will need to do a lot of research and calibration and use a good oscilloscope.
In porous media, the speed of sound is not related to density in a simple way as in homogeneous media. But you could do some experiments to calibrate your device for "good dough" and "bad dough" by using some appropriate parameter.
If you plan to have a handheld device with just a color coded output in the end, you will need to design your own signal processing hardware and software, right?

 

[sophiecentaur]

Using ultrasound is hardly something that will count as "easily do" in this case.
I mean, to get something meaningful you will need to do a lot of research and calibration and use a good oscilloscope.
In porous media, the speed of sound is not related to density in a simple way as in homogeneous media. But you could do some experiments to calibrate your device for "good dough" and "bad dough" by using some appropriate parameter.
If you plan to have a handheld device with just a color coded output in the end, you will need to design your own signal processing hardware and software, right?

You are right; it isn't an easy peasy project to start ones amateur constructor career but it doesn't actually involve any technology that's cutting edge so not beyond the scope of someone with a bit of experience. As you say, calibration is the name of the game - both in the development and for every batch of dough (before and after measurements each time, I think). The electronics and software are really the easy part of the project, I think. It's the transducer and the measurement probe that will be the most demanding to get right.

 

[nasu]

You are right; it isn't an easy peasy project to start ones amateur constructor career but it doesn't actually involve any technology that's cutting edge so not beyond the scope of someone with a bit of experience. As you say, calibration is the name of the game - both in the development and for every batch of dough (before and after measurements each time, I think). The electronics and software are really the easy part of the project, I think. It's the transducer and the measurement probe that will be the most demanding to get right.

I did not say that is cutting edge technology. It's not easy to make a good clay pot without experience. And is thousands of years old technology.:)
There are two issues here.
The first one is to find a good enough parameter or thing to measure so that is strongly correlated with the quality you are trying to monitor (here "good" dough).
It's not obvious that this parameter will be found. People were (and are) trying to use ultrasound to measure bone density (instead of x-rays) for many years. But the results are still not good enough to displace the standard x rays even though there is strong pressure to eliminate radiation wherever possible.
Of course, what is not good enough for health industry may be good enough for pizza.

Second, he said he would like to have a self-contained device with some simple output and not some oscilloscope curves. This can be done of course even by amateurs, with these new "microcomputers" (Gumstix for example) available for a few hundred dollars. But it's hardly something "easy" unless you did it before and have some experience.

The transducers should not be a problem as I understand he plans to use low frequency ultrasound, like the one used for measuring distances or location for visually impaired. They are cheap and come with controllers.

 

[sophiecentaur]

I don't like the idea of measuring distance much because that only gives you size or time and not both (i.e. how would you know the speed if you didn't know how big the balls were?). What I would do would be to try to look at reflection coefficient. That depends just on reflection at an interface and would involve measuring two amplitudes - the amplitude when the main pulse enters the probe and the amplitude when the reflected pulse turns up. The time between those pulses would always be the same (just length of probe).
You don't need to spend hundreds of quids for a processor An Arduino, which has a whole development community and loads of interfaces, is about 30GBP. Programming it is very easy. A video ADC would do the pulse measurement for you easily at 28kHz. By no means trivial though. It depends on ability and available time, as with all these investigations. Ultrasound is used in food prep, as the few google hits tell us.

 

[nasu]

I don't like the idea of measuring distance much because that only gives you size or time and not both (i.e. how would you know the speed if you didn't know how big the balls were?). What I would do would be to try to look at reflection coefficient. That depends just on reflection at an interface and would involve measuring two amplitudes - the amplitude when the main pulse enters the probe and the amplitude when the reflected pulse turns up. The time between those pulses would always be the same (just length of probe).
You don't need to spend hundreds of quids for a processor An Arduino, which has a whole development community and loads of interfaces, is about 30GBP. Programming it is very easy. A video ADC would do the pulse measurement for you easily at 28kHz. By no means trivial though. It depends on ability and available time, as with all these investigations. Ultrasound is used in food prep, as the few google hits tell us.

Who said to measure distance? I only said he can use these transducers (used for locating obstacles) for whatever he wants to. By transducers i mean the devices used to produce and detect ultrasound. I am not sure if you mean the same thing by "probe".
Actually this is the whole point of not being an easy problem. To find out what to measure. I did not suggest at all to measure some size of the sample. It would imply that you know the speed of sound in the medium and this is not the case at all.

The amplitude of the wave reflected by the surface may indeed be correlated to the quality of the dough. You just have to make sure the amplitude of the incident wave is always the same. Or at least to have a way to measure it every time.

 

[sophiecentaur]

The transducers that people buy for distance measurement use pulses. There are also high power CW transducers for 'physical' stuff like cleaning. I am assuming a pulse type.
I was making general assumptions about knowledge of measuring things like density, using sound. You can look at speed directly (timing) or by reflection coefficient at a boundary with a known substance. ~~~~A timing measurement needs either to know distance to get speed (connects to density). A reflection coefficient method doesn't require timing but does require a comparison of amplitudes (forward and reflected). Anything that can measure one amplitude will measure the other and allow a ratio to be calculated. A suitable 'probe' would consist of a length (rod) of a medium with a suitable density and low loss with a transducer at one end and a measuring transducer at some other point on the rod. The other end of the rod would be in contact with the dough. The amplitude ratio would give an estimate of density but for a better estimate, the unrisen dough could also be measured and the proportional increase in density could then be inferred. That would be better. Clearly, there would be calibration runs with 'before and after ' measurements for various dough mixes etc etc.
As I said earlier, a video ADC could give a good picture of the bursts of ultrasound and could produce, with the help of some memory, possibly, a recording of the first and second blips which a processor could easily analyse and produce the required numbers. We should bear in mind here that the user would be an experienced baker whose dough would already be pretty consistent and who is only after an optimum condition within an already pretty well controlled process.
The 'probe' would need a fair bit of development in order to match the transducers to the rod material to get a good signal to noise ratio for the signal to be digitised. That would be the only unknown quantity in the development.

 

[nasu]

OK, i suspected that what you called "probe" may be referring to something else than the meaning we usually use in our lab (the transducer itself).
This setup may work, sure. Assuming that there is a good and reproducible contact between the end of the probe and the dough. Using some contact gel as in ultrasound imaging may not be a allowed. :) Otherwise the reflection will depend more on the amount and distribution of air at the interface between the two.Which is OK if that won't change with the evolution of the dough.

 

[sophiecentaur]

The "probe" would consist of a rod of suitable material ( speed similar to dough, ideally) with a transmitter one end and a receiver at some point. Yes, a good contact with the dough would eliminate bad reflections at the interface. As long as the layer of air is thin, the intermediate reflection would not affect things too much. I have to admit that the wavelength of the ultrasound in dough is important because there is resolving power (pulse width) on one hand and availability of transducers on the other. I wonder whether a lower frequency could be better.
Looking back on using Time Domain Reflectometers, I have to conclude that a step function could be best for this purpose. Not easy to obtain with a pressure output transducer.

 

[nasu]

The "probe" would consist of a rod of suitable material ( speed similar to dough, ideally) with a transmitter one end and a receiver at some point. Yes, a good contact with the dough would eliminate bad reflections at the interface. As long as the layer of air is thin, the intermediate reflection would not affect things too much. I have to admit that the wavelength of the ultrasound in dough is important because there is resolving power (pulse width) on one hand and availability of transducers on the other. I wonder whether a lower frequency could be better.
Looking back on using Time Domain Reflectometers, I have to conclude that a step function could be best for this purpose. Not easy to obtain with a pressure output transducer.

Unfortunately it will. Otherwise they won't use gel in medical ultrasound. You will get a strong reflection from the air and almost no power transmitted to the second medium.
For imaging may even be less critical but if you plan to use reflectivity as a quantitative indicator I think that air will be big trouble. Even thin.
But again, maybe at low frequencies (kHz) the air will matter less. My experience is with frequencies over 2 MHz.

 

[sophiecentaur]

Unfortunately it will. Otherwise they won't use gel in medical ultrasound. You will get a strong reflection from the air and almost no power transmitted to the second medium.
For imaging may even be less critical but if you plan to use reflectivity as a quantitative indicator I think that air will be big trouble. Even thin.
But again, maybe at low frequencies (kHz) the air will matter less. My experience is with frequencies over 2 MHz.

I agree that the step change in refractive index, at an air gap would have an effect but this system does not need to involve wavelengths that are small compared with a gap.
Having re-thought this, with some actual figures, I realize that the wavelength in the dough, for a cheaply available 40kHz transducer is likely to be a few cm. That is not short enough do make any measurement of signals in a dough ball of similar size. It calls for several hundred kHz, I think. That is harder and more costly. Medical diagnostic ultrasound works with frequencies in the MHz region for this reason.
It may be a non-starter.