Enlarger-controller features you've never seen

In this posting, I describe some new time-saving things one can do with test strips or an easel meter, with the assistance of a smart enlarger-controller. I have not seen these features described or implemented anywhere, so I wanted to post them to prevent anybody from patenting them. These ideas are suitable for both tungsten- and LED-based enlargers. People like Derek (@dkonigs) intend to put products into production, and I don't want them blocked by patents.

Background

I have designed and built two LED controllers. Here is the second one:



There are six "soft" buttons below the display. The upper row of three buttons correspond to the items in the upper row of the display. Likewise, the lower row of buttons are for the lower row of the display. The two columns of buttons on the left and right sides are "hard" in that each performs a single function. In my first controller, the upper row of buttons was above the display, putting them closer to the display-items they affect, but I found that when pushing them, my hand partially blocked the display. So I put all buttons under the display, which has the second advantage that one can put a grayscale above the display like the RH Designs Analyser Pro. If your controller has fewer functions, you can replace the soft buttons with hard buttons.

My display is character-based. A graphical display would look better, and could display a grayscale so it wouldn't need to be a fixed part of the control panel.

The left knob is marked "Numeric" because it's used to enter numbers. Pressing down on the knob while turning it causes the number to change in larger (5x) steps. The right knob controls the display's brightness, and pressing it turns off both the display and safelight in order to blackout the darkroom for color work.

For a LED-based lamp with separate color-control, one can have a hard button on the controller labeled "Red", as I do in the photo above. It turns on the red LEDs, letting you position a tool before starting a burn or dodge. During exposure, the red LEDs stay on to help maintain position.

Test Strips

Test strips can do more than tell you exposure time. One test strip can determine both exposure and grade. And if its range is large, it can also determine burn/dodge times. Here's a sample display showing determination of exposure and grade:



In this display, we are telling the controller, "On step 1 of the test strip, zone 3 is correct, and on step 3, zone 6 is correct." By "zone X is correct", we mean that tone X on the step should print as tone X. The zones are marked by the numeric scale under the arrows, but as shown in this picture, the tones can also be marked using a grayscale. Thus, zone terminology is not needed when using a grayscale.

The positions of the arrows and the step numbers next to them are set using the knob after selecting them with the soft buttons. When changing step numbers, they change in quarter-step increment, 1 to 1-1/4 to 1-1/2 to 1-3/4 to 2 and so on. When increments are large, tones are seldom correct on a step, and this quarter-stepping lets you estimate where the tone would have been correct. "Let's see, this skin tone would be correct a little before step 6, so I'll enter step 5-3/4."

After entering the two steps and their zones (or positions on the grayscale), the controller calculates both exposure and grade. In the above example, the grade is 1.3.

Also, you can enter one step and zone, omitting the second, and the controller will calculate the exposure, leaving grade unchanged.

Here is a more general way of specifying steps and tones.



The upper row of the display is telling the controller that "On step 2, elements in zone 3 shall print at zone 2.5." The lower row is saying, "On step 3, elements in zone 7 shall print at zone 6." The controller is telling us that doing so will print at grade 1.4. Instead of zone numbers, the upper row could use arrows or letters (such as "S" and "P") at the correct positions under the grayscale. "S" would stand for "step" and "P" would stand for "print". Likewise, the lower row can have two such points on a second grayscale. Instead of having two grayscales, you could put one grayscale in the middle of a graphical display, with a pair of points above it (1st row), and a pair under it (2nd row).

Dual-scale 1: If the controller has a graphical display, it could show two grayscales. The upper one would always be the same: black at one end and white at the other. The grayscale under it would be horizontally shifted and/or scaled, showing how the tones on a given step of a test strip will appear after the exposure/grade changes described above have been made. For any point on the upper scale, the same location on the lower scale shows how that step-tone will print.

Dual-scale 2: Furthermore, if the user made a low contrast test-print (not test strip), it could be represented by the upper grayscale, and the final print could be represented by the lower grayscale. He could then use the knob to shift and/or scale the lower grayscale to map print-tones from the test print as he wishes, thus determining exposure and grade.

A test strip can also determine burns and dodges. Here's a sample display for this feature:



We are telling the controller, "I want zone 9 on step 7 to print at zone 8," and the controller is telling us that requires a burn of 1.3 stops. This example illustrates burning in sky. Instead of numeric zones on the display, one could display two points on a grayscale as described above.

All of us are familiar with test strips based on changing exposure time. But test strips can also be based on changing grade. This feature is more suitable for LED-based lamps with separate control of green and blue because there is no filter-swapping. In fact, for split-grade work, a test strip could hold one color (green or blue) constant while changing exposure of the other color.

Easel Meter

I recommend the easel meter sold by Darkroom Automation (darkroomautomation.com). For your controller, you could also design your own meter that transfers readings directly into the controller, eliminating numeric entry. In either case, similar to steps on a test strip, two meter readings will let the controller compute both exposure and grade. Here's a display for it:



In this display, we are telling the controller, "Reading A shall print at zone 2 and reading B shall print at zone 6." The controller is telling us the grade will be 3.6. The positions and meter readings of A and B are selected using soft buttons and are changed with the knob. A and B can be displayed as points on a numeric zone scale or on a grayscale; the photo above shows both.

Note that the above order is the opposite of the Analyser Pro. In it, you enter the exposure time and grade, and it positions the points on the grayscale. But above, you enter the positions (tones) you want for the meter readings, and the controller tells you time and grade. I prefer the second approach because you are directly telling the controller what you want instead of manipulating settings to yield what you want.

The point "I" in the display is for information. When a meter reading is entered for "I", the "I" moves to the corresponding tone it will print at, allowing you to check the print-tones of various elements in the image on the easel. If you then move the "I" with the knob, the controller shows the burn or dodge needed to print that meter reading at that tone.

The same capabilities are portrayed in the following display that does not use a grayscale or zone scale:



The upper row is saying that "Meter reading 4.3 shall print at zone 2, and reading 2.6 at zone 6." The controller tells us this will be grade 3.6. The lower row lets you enter a meter reading, and the controller tells you the zone it will print at, allowing you to check the print-tones of various elements in the image on the easel. Pressing the soft button for "BurnDodge" lets you enter a meter reading and zone, and the controller will tell you the burn or dodge needed.

When using a meter, the print can differ from what you want due to a color cast in the negative, or batch-to-batch variation in the paper or developer, or partially exhausted developer. You can correct for these anomalies by making a test print and entering the following information:



In the upper row of this display, we are telling the controller that what should have been zone 3 printed at zone 3-1/4, and 6 at 6-1/2. The lower row shows the green and blue biases computed by the controller. It's telling us that green will be boosted by 0.4 stops, and blue will remain unchanged. You can also enter these bias numbers yourself. Instead of four numeric zones, four points on a grayscale could be shown instead.

For a LED-based lamp, the green and blue LEDs can be changed by the biases. For a tungsten lamp, the biases will enter the calculation of time and grade.

Other Ideas

Split-grade. We all love split-grade printing. In theory anyway. The display below breaks up a burn or dodge into green and blue components that will result in the same tone-change.



In this example, we are telling the controller that we want to burn 0.2 stops in a zone 7 area. The controller tells us that is equivalent to a 0.3 stop green-only burn or a 0.6-stop blue-only burn. The amount of burn/dodge could have come from either a test strip or meter readings. This feature makes it easy to create dramatic clouds. When using contrast filters, green is your 00 or 0 filter, and blue is your 5 filter.

Flashing. If you flash the print, the controller can calculate the exposure adjustment needed in order to print an element at a desired tone. A display would let you enter the degree of flashing and the desired print zone. The controller would tell you the zone to specify for that element when printing that will cause it to print at the desired zone.

Vibration. The controller can delay a settable number of seconds before starting exposure, so that vibration will die out.

Changing height. When changing height, measure the before- and after-height of the lens, and enter them into a display. Any units (mm, cm, inches) will work. The controller will update the exposure based on the inverse-square law.

* * *

I hope this posting inspires somebody to implement one or some of these useful ideas, and prevents anyone from patenting them.

Mark

Thank you for this! It’s quite amazing! I think the only way to stop someone from just patenting your clever ideas is to patent them yourself and then granting rights to whoever wants to implement them.
Do you have any desire to mass producing your unit? I’d love to use one of those!

Public disclosure goes a long way to preclude patenting an idea. In the old days you could just file an application to keep patents away from an idea - filing was only $50. And applications were 'secret,' allowing you to stamp 'patent pending' on your equipment in the hopes of preventing someone copying the design and nobody knew what part was patented.

As for the practicality of patents, the size of the market for darkroom equipment wouldn't support the fees for lawyers, filing and patent maintenance.

None of the old-line firms are introducing new timer/controllers, it's all private enterprise where the owner of the firm is also, as they used to say, 'Chief Bottle Washer and Dog Walker.'

If you don't have an LED head there are units that have much of the functionality, though with much simpler displays, available from Darkroom Automation & RH Designs.


Disclaimer: I am the Chief BW & DW at Darkroom Automation. Gotta go, the dog is whining at the door...

Nice functionality, however, I suspect if it is Arduino based, most people would program the display and buttons to their own preferences.

I am curious about the front plate. How did you get the letters on it and how did you cut the holes and countersink the screw holes? It looks very nice !!!!

I am curious as to why so many wires come in the back. Is it really just a controller for the LED drivers that are somewhere else? In that case couldn't it just have a single serial communication link to the drivers?

For example the I like working with my existing controllers (for dichroic filter heads) and would probably copy them. They are feedback loop controllers and only have a single cable to attach to the head.

For prototype panels and cases try:


Very high quality, fast service and reasonable prices. I use them for client work. Located in Nova Scotia of all places.

Pre-prototype front panel overlays can be made with artwork printed to RC paper that is then sprayed with Krylon. For slightly less rugged applications an ink-jet print on heavy glossy paper and sprayed with Krylon is perfectly adequate.

DIY 'bathtub' circuit boards with the front panel outline, holes and cut outs as copper traces can be used as the underlay. It gives you precise lines for filing cut outs and precise centers for drilling. I make them as negatives - the boards are all copper foil that is only etched at centers cut out edges; etching goes faster and I don't use up the etchant any more than necessary.

All that is now superceded by using inexpensive laser engravers for the artwork. Get one powerful enough and it can blow holes in the aluminum. But even cheap ones can cut through thin plastic or birch plywood. A wood front panel ... that's a thought.

Fantastic link, thanks!

Nicholas Lindan said:

Public disclosure goes a long way to preclude patenting an idea. In the old days you could just file an application to keep patents away from an idea - filing was only $50. And applications were 'secret,' allowing you to stamp 'patent pending' on your equipment in the hopes of preventing someone copying the design and nobody knew what part was patented.

As for the practicality of patents, the size of the market for darkroom equipment wouldn't support the fees for lawyers, filing and patent maintenance.

Click to expand...

Another great thing about the darkroom equipment market today is that is relatively old and mature. That means that there are a ton of patents covering everything about it, but the majority of those patents are already long since expired. Which means that you don't need to worry about violating them, and can even use them as a source of information for how to do things.

So even if a product brags about being covered by patents on their seldom-updated website/manual, if you dig in and find the record of those patents, you may very well find that they're no longer in effect.

Nicholas Lindan said:

For prototype panels and cases try:

Protocase | Custom Manufacturing with Fast 2-3 Day Lead Times | Protocase

Protocase

www.protocase.com

Very high quality, fast service and reasonable prices. I use them for client work. Located in Nova Scotia of all places.

Click to expand...

For my first Printalyzer Timer prototype, I actually used a different company: Front Panel Express
Their quality was also quite nice. But the problem is that they get kinda expensive kinda quickly.

For my next prototype, I'm seriously considering switching over to Protocase. I've even done some initial design mockups in their software, to see how well it would work. At low/single quantity, the price difference is honestly negligible. But once you want to make 100+ of something, it can be significant. (Also, Protocase admits to their quantity discounts while FPE only publishes very modest percentages that don't amount to much.)

ic-racer said:

Nice functionality, however, I suspect if it is Arduino based, most people would program the display and buttons to their own preferences.
I am curious about the front plate. How did you get the letters on it and how did you cut the holes and countersink the screw holes? It looks very nice !!!!
I am curious as to why so many wires come in the back. Is it really just a controller for the LED drivers that are somewhere else? In that case couldn't it just have a single serial communication link to the drivers?
For example the I like working with my existing controllers (for dichroic filter heads) and would probably copy them. They are feedback loop controllers and only have a single cable to attach to the head.

Click to expand...

I used Front Panel Express. And as Derek mentioned, they are expensive. That panel cost me $130. Ouch! Based on that, I estimate that Derek's Printalyzer prototype case cost $600-700.

My first controller used the Arduino Uno. The second uses the $20 eval board for the PIC18F57Q84, another 8-bit microcontroller, which worked out well because it's a more modern processor with more flash and four 16-bit PWMs. If I had to do this again, I would look at the Raspberry Pi Pico or other 32-bit microcontrollers.

Here are the wires out the back, from left to right:


I couldn't find any reasonably priced high-power connectors suitable for the RGB lines, so I just used three RCA connectors. On the other hand, if the high-power LED-drivers were to be moved out of my controller and into the lamp, then a VGA connector could be used for the low-power PWM signals and the flasher LED.

Regarding patents: The products produced by RH Designs are dated and in need of a refresh, and my concern is they might patent new ideas as part of their update. They patented the grayscale, so they might try to patent something new.

Mark

albada said:

I used Front Panel Express. And as Derek mentioned, they are expensive. That panel cost me $130. Ouch! Based on that, I estimate that Derek's Printalyzer prototype case cost $600-700.

Click to expand...

With tax and shipping, it actually only came to ~$470, but that's still a lot. (And that's just the enclosure, not anything that went inside it.)
Of course their prices have probably gone up since then.

Regardless, while acceptable for a prototype, the price needs to be a lot cheaper in quantity for it to make sense for any sort of product. I'm just glad that Protocase is far more open about their quantity discounts, and once I'm trying to build 100+ they give reasonable-enough pricing.

Another big gotcha to look out for, is the sheer number of value-add things that drive up the cost. Like wanting a nicer finish, or some sort of surface treatment, or better quality printing, or really anything beyond "cut the metal"... and it adds up fast.

Beautiful, but I couldn't remember how to use it. I want a setup with huge black phenolic knows. Like a variac .

Knobs

mshchem said:

Beautiful, but I couldn't remember how to use it. I want a setup with huge black phenolic knobs. Like a variac .

Click to expand...

My Variac is 3 feet from me. No kidding.

Here's an idea for the front panel of a simple controller for a LED-based enlarger:


That's four big knobs to set everything. The preserved-zone knob let's you keep the highlights unchanged when you change grade to get the shadows right. Or the opposite: preserve shadows while setting grade to make highlights correct. In general, when changing grade, some tone will remain unchanged. This knob selects that preserved tone. The toggle switches make split-grade printing easy.

EDIT:
For color, you could add a toggle switch for MONO or COLOR. In color mode, the time knob stays the same, but the other three knobs become red, green, blue.

Mark

albada said:

My Variac is 3 feet from me. No kidding.

Here's an idea for the front panel of a simple controller for a LED-based enlarger:

Knob - Exposure time.​

Knob - Brightness.​

Knob - Grade.​

Knob - Preserved zone that is unchanged when changing grade.​

​

Display - Shows the above four numbers.​

​

Button - WHITE (for focusing and composing).​

Button - RED (for positioning a tool for burn/dodge).​

Button - START (or tap a foot-switch)​

​

Toggle switch - Enable/disable red LED.​

Toggle switch - Enable/disable green LED.​

Toggle switch - Enable/disable blue LED.​

That's four big knobs to set everything. The preserved-zone knob let's you keep the highlights unchanged when you change grade to get the shadows right. Or the opposite: preserve shadows while setting grade to make highlights correct. In general, when changing grade, some tone will remain unchanged. This knob selects that preserved tone. The toggle switches make split-grade printing easy.

EDIT:
For color, you could add a toggle switch for MONO or COLOR. In color mode, the time knob stays the same, but the other three knobs become red, green, blue.

Mark

Click to expand...

Even if the knobs are like the radio on my Subaru, not real knobs, some sort of digital thing, still easy to use.

I have a couple of the late, Calumet era, Zone VI VC cold light heads. All knobs, except I use Metrolux II and RH Designs Vario controllers, original Zone VI compensating timer is knobs.

I have digital cameras that everytime I pick them up I need a review of the instructions.

I'm ready to buy

Also, toggle switches would be Awesome! Really amazing work!

The RH patent has long lapsed. Along with all the $15,000 expense of getting a patent you then have to pay $4,000/year to maintain it. RH stopped paying maintenance and the patent lapsed in the US.

As I understand it the owner/founder of RH designs sold the company some years back. I don't know if the present owner is just selling off the old stock, producing new stock or actively pursuing new product development.

I don't think anyone working in this area has much to fear from litigation, there isn't enough money in it to make it worth anyone's while.

Nicholas Lindan said:

The RH patent has long lapsed. Along with all the $15,000 expense of getting a patent you then have to pay $4,000/year to maintain it. RH stopped paying maintenance and the patent lapsed in the US.

As I understand it the owner/founder of RH designs sold the company some years back. I don't know if the present owner is just selling off the old stock, producing new stock or actively pursuing new product development.

I don't think anyone working in this area has much to fear from litigation, there isn't enough money in it to make it worth anyone's while.

Click to expand...

Here's the fee schedule for the US Patent Office: USPTO Fee Schedule

For an individual person, which they call a "micro entity", the submission fee is only $44.00. Add $300 for search and exam, and another $240 issuance fee. That's $584. Double that for a "small entity", which I suspect means a small company. I can envision that somebody would be willing to pay these amounts to eliminate a likely future competitor, even with our low sales volumes.

The maintenance fees are $400 (micro entity) and $800 (small entity) at 3-1/2 years. Fees are higher in later years. Where did those $15,000 and $4,000/year numbers come from?

Mark

Here's another practical idea of using a densitometer connected to the controller.

On a step in a test strip, measure the density of an element with correct tone. That measurement will determine exposure time. If a second measurement is taken of another element, it will determine grade.
In general, in my OP, any mention of a zone on a test strip or print can be replaced with a densitometer measurement. The user can also be given the option to say that a given measured tone should be a little lighter or darker.

In my OP, I mentioned making a low contrast print, and specifying two zones of elements in it to determine time and grade. If those two zones were densitometer measurements, we have a simple method of determining time and grade that will be new to all of us:


This idea replaces test strips with a single low-contrast print.

EDIT: The user could measure a third point, and then move it on the grayscale. The controller would indicate how much burn/dodge is needed to cause that tone-change.

Mark

albada said:

1. Guess time, and make a grade-0 print.​

2. Measure two spots on the print with the densitometer. The controller will display those two spots on the grayscale.​

3. The user can move those two spots to their desired densities, and the controller will calculate time and grade. Or, like the Analyser Pro, the user could change time and grade, causing the spots to move to their desired densities.​

Click to expand...

If you know the paper's sensitivity, and can measure the light, you can skip step 1. If you can do steps 2 and 3, you already have the tools needed to accomplish this. That's how the Analyser Pro works.

dkonigs said:

If you know the paper's sensitivity, and can measure the light, you can skip step 1. If you can do steps 2 and 3, you already have the tools needed to accomplish this. That's how the Analyser Pro works.

Click to expand...

Yes indeed. If you can measure the light, you can measure important spots on the easel, and there would be no need to make that print I describe above.

For some background, I've discovered that an easel meter becomes inaccurate if the negative has a blue color-cast as occurs with some developers and fixers. PE (Ron Mowrey) mentioned this problem, and I've seen it with Ilford negatives: They print with a bit of extra contrast, causing target densities to be a little off. Also, some negs have a pink/magenta cast from left-over dye. Batch-to-batch variations are also a problem. That's why I added the Bias feature to my controller. But test strips automatically adjust for these things, which is why I paid much attention to them.

An easel meter would be more accurate for black-and-white work if it could measure the spectral composition of the light, compare it to the spectral composition of the calibration-light, and compensate for the resulting contrast-shift. That would eliminate the problem of color-cast in negatives, which has been frustrating for me. I could do that in the following clumsy way with my LED head: Through a negative, measure the red-green-blue components separately by only turning on those LEDs. The controller could then compute the exposure- and contrast-correction (or biases) based on the red-green ratios and blue-green ratios for the negative and calibration. Luckily, that clumsiness would need to be done only once per roll.

Mark

DIY'ers and perhaps Derek will be interested in this: The plastic console case I used is made by Bud Industries, model Plasticase PC-11491. It's the perfect size for an enlarger controller, and the sloped top panel is helpful and attractive. Their website is https://www.budind.com/, and this console case is here. If you click on "Check Stock/Pricing", you'll see prices and stock of their retailers. The typical price is US$20-22 for one, and $13-14 for 50 or 100. Those prices are attractive. But the problem with this case is its rear.



The panel at the lower left is not large enough to accommodate the three outlets we need for mains, enlarger, and safelight. So you must cut plastic as I did on the right, or perhaps cut a large opening and mount an aluminum plate across part of the back with the outlets on it. Either way is kludgey.

The case has two parts: upper and lower. I put all the low power and digital stuff in the upper half, attached to the aluminum top panel. As an inspiration (or perhaps a fright) to the DIYers, here it is:



The ribbon cable brings PWM signals (and a few others) to the high-power electronics in the bottom part of the case, shown below.



The three big chips along the left side are the high-power LED drivers by Mean Well. The vertical "chip" below them is the relay for the safelight.

As you can see in both the top and bottom, I assembled few individual electronic components. Most of the work was hooking chips and components together. Wiring. Mostly wiring. That is not hard work, so I encourage folks to create such DIY projects. If you're interested in learning how to build such electronics, a good place to start is Arduino and its website here. They cater to beginners, so you'll find clear instructions and a community at that site. My first controller was an Arduino Uno, but the latest one used the PIC18 (the red circuit board on top) because I needed more capability and wanted to try something different.

For the the upper and lower circuits, I used perf boards, which are those green boards full of holes (perforations). A producible design would have put all the components, including buttons and such, on one large circuit board under the control panel on top. That's what @dkonigs is doing with his Printalyzer. Maybe I'll design a circuit board someday.

I might as well finish with a helpful tip for other DIYers. Notice the two thick copper wires near the lower-left corner of this photo of the bottom of my high-power board:



These are house wires! Each provides a low impedance path to the power supply so the noise from a chip that periodically draws current won't affect other chips. You still need bypass caps. A second purpose for the thick wires is to transfer strikes of static electricity (ESD) away from the circuitry. I call these two wires the "power and ground rails". In a circuit board, there are usually multiple planes of "printed" wires, and often, an entire plane is devoted to ground and another to power. These ground planes and power planes serve the same purpose as my rails. I asked an experienced EE at work about my rail approach, and he agreed with it. In fact, he said he's done similar things in the past.

A bad design would be to daisy-chain the ground and/or power wires among chips. An EE did that at work many years ago, and I had write the firmware to get his circuit board working. That was an adventure. The company fired him, and had a competent EE take over the design. And I learned a lesson: Make certain your ground and power rails/planes are solid!

A final clue: Learn about TVS diodes/arrays, and use them to protect your circuit from ESD strikes. Those diodes shunt those zaps onto the ground/power rails where they won't hurt anything.

BTW, I am a software geek and not a EE, so EE's here are encouraged to correct me.

Mark

Pretty impressive work!

albada said:

DIY'ers and perhaps Derek will be interested in this: The plastic console case I used is made by Bud Industries, model Plasticase PC-11491. It's the perfect size for an enlarger controller, and the sloped top panel is helpful and attractive. Their website is https://www.budind.com/, and this console case is here. If you click on "Check Stock/Pricing", you'll see prices and stock of their retailers. The typical price is US$20-22 for one, and $13-14 for 50 or 100. Those prices are attractive. But the problem with this case is its rear.

Click to expand...

While an off-the-shelf case may have its benefits, by the time you've made the necessary custom top and rear panels you haven't saved quite as much as you think.

But as far as designs in general, I know the "sloped console" one does look nice on the surface. But it has one big problem from an assembly/internal-design point of view. You can't use the circuit board from the top panel to mount any of the front or rear connectors. So if any of your connectors are board-mounted, that means a minimum of 3 PCBs to make it work with front and rear jacks.

The first prototype of the Printalyzer Enlarging timer actually attempts to fit everything (panel, rear connectors, front ports) onto a single PCB. Of course now that I'm getting started on the second prototype, I've decided to split off the rear ports into their own PCB. Two real reasons for that... 1) To isolate everything that has to touch mains AC into its own board I can update/finalize independently. 2) Because it actually lets me reduce the case depth a bit.

albada said:

A final clue: Learn about TVS diodes/arrays, and use them to protect your circuit from ESD strikes. Those diodes shunt those zaps onto the ground/power rails where they won't hurt anything.

Click to expand...

I've definitely gone overboard with TVS diodes in my designs, just to be safe. The most frustrating part with them, is actually finding single-line TVS diodes that aren't damn near microscopic. I'm now on the 3rd choice of diode with my various densitometer project PCBs, each time increasing the size a little, and they still look like specs of dirt when you look at them on the table without a microscope.