Jonathan's Notes

  • Jonathan Brodsky - Fri 28 March 2025 - sculpting, miniatures, casting

    I've recently improved my casting setup at home a little bit - I'm now using a jewelry mold vulcanizer for making my molds when possible. Its normally used by jewelry makers for creating molds for doing wax multiples, to support investment casting. Luckily the mold material supports low temp metals as well, so I can just dump pewter into it.

    absolutely beat to shit bit of orc tech

    There are a bunch of these machines kicking around, since they seem to be quite indestructible, and they are also still being manufactured. It lowers my per mold cost, improves my process speed (1 hour for a mold instead of 48), and I get better results due to not having any bubbles.

    I'm still doing gravity casting into the molds however, which seems to be the final hurdle in terms of quality. I've been looking around for a while a different solutions that folks have. This video by zombie smith miniatures (the guy that designed and sculpted quar) was the first thing I found. It seems to have quite a bit of setup costs, particularly the fact that he is using 9in molds for casting. This means that he had to build his own mold boxes. There are a few places to buy mold boxes still (contenti and spinbox both sell some that appear that you could chuck them in an oven).

    The second thing I found was this article from 1966 showing how to build a 4in spincasting rig. The author is using RTV silicone, which I'd like to stay away from if possible. But the idea of mounting it on some simple off the shelf motor is pretty interesting.

    This one using a washing machine as the motor and mount is pretty interesting - I'm not really sure how much the author is getting out of this setup. Maybe a bit, since you need to make your own pulley system to gear down an off the shelf motor if you go that way.

    The final thing I found is someone from reddit that had built their own. No links because I've just been chatting with them, but they are doing a few interesting things. First, instead of making round molds, they are just doing little square molds like I'm using. The are also using something that is incredibly easy to build.

    Diagram showing the casting setup. Description follows

    The idea here is that you have a bar with a central pivot, at the end of the bar are little carriers for your molds. These are attached to the bar via hinges, so that when the bar comes up to speed, they tilt up.

    So to use this, you pour the molds with the whole thing stopped, then spin it up to speed (by hand! motor not necessary, though it is a nice process improvement), and keep it spinning there until your casts solidify. I am super into this solution. And the results they were getting were amazing. So I'm going to build one. I really like the fact that it can be done with mostly stuff from the hardware store.


  • Jonathan Brodsky - Fri 21 February 2025 - sculpting, miniatures

    Mossy ent striding towards the camera. It has a giant hand Mossy ent striding away the camera. Mossy ent being threatened by a group of goblins

    I sculpted a big treeman for entmoot25, an instagram sculpting jam organized by # @whitetigertablecraft, # @michaelcwiatrak_art, and kyle / # @honkteam. I'm still having fun getting bigger and bigger with polymer clay, so figured this was another good one to do. I think most folks doing the challenge glued together a heap of twigs, but I strangely found this difficult to pull off. I'm not sure exactly why, maybe I was using the wrong epoxy or something. There is a pretty sad ent kicking around the house from this initial pass. Luckily the deadline got extended, so I had more time to work on it.

    I initially setup two armatures to sculpt onto, but ended up just focusing on one. The nice thing about ents is that the don't particularly need reasonable proportions or even body symmetry to read right.

    I think I may end up going back to that other sculpt at some point, I really like its pose a bunch.

    two grey ents on a board photo showing more of the detail in polymer clay of the ent I finished Comment on mastodon or bsky


  • Jonathan Brodsky - Wed 12 February 2025 - sculpting, miniatures

    Big painted goblin threatening with raised fist

    Its been a bit of a goblin year for me, relatively speaking. I went from an average of less than one goblin, to probably 50 goblins either sculpted or drawn in the year. I did goblin week last year early on in my sculpting, and I didn't know what an "official" goblin looked like. Not that there is an official goblin, but they def weren't reflecting the big nose, floppy elf ears, tiny torso look that most goblin art has.

    For this year, I wanted to try out making a posable multipart goblin, so I could make a little army. I also decided to work quite big, and use only polymer clay for the sculpting, rather than my standard epoxy.

    goblin rouged out without legs goblin with a hat, looking saucy cat sniffing goblin

    For casting this time, I make some sprue blobs, the idea being that I can get a better resin pour if I have more resin on top of the mold. Then I can reuse this extra nothing junk as filler for diorama & terrain bits.

    goblin with his arms and various chunks of stuff

    The casts came out alright, but I was hoping the various pastel colors would be more shocking, and would look cool in a lineup of figures. I ended up deciding to just paint them like standard. I'd like to mess around with different resins and see if I can get a more toy like finish at some point.

    Goblins in various pastel colors

    Here is the king goblin! he's quite tall, around 65mm, and is set on a 40mm square base.

    Side view of the finished painted goblin

    I also did some bonus gobs for the week - I'm very happy with this proper looking guy - I tried to follow some reference goblins for proportions closely, and I think I got it alright. The hands are starting to come out how I want them to as well.

    tiny gob with a giant spear

    I also made a quick linocut postcard of a screaming goblin. I've been sketching a bunch, and this one stood out as something that would be fun to make multiples of. I added some watercolor - I'm super into the work that the still tower does, and I thought I'd give it a try. Not quite as successful, but I haven't done much lino cutting in my life, so I'm being gentle with myself.

    I managed to give a few away, but I still have a few more - if you'd like one let me know your address on a dm or email, and I'll mail it to you.

    Yelling Goblin lino cut print. It has green flesh and a blue helmet

    Comment on mastodon or bsky


  • Jonathan Brodsky - Tue 13 February 2024 - sculpting, miniatures

    I wrote these notes approximately a month into sculpting miniatures, with the hope that I could get some other people fast forwarded down the path. If you jump to the end, I've got links to folks that are doing a great job of tutorializing via videos.

    With a year of sculpting miniatures hindsight, I might approach things different. Maybe I'll update this with better notes someday.

    materials

    some kind of wire to make armatures - I'm using 1mm steel galvanized stuff that I got at the hardware store. Some people use paperclips, but I found they break very easily. The important thing is that you want your armature to be very stable. I was initially using 0.5mm wire and my models were flexing all over the place. The sculpting media is pretty thick, so you gotta use some serious force to get it on the model.

    you will need at a minimum, green stuff - this is sold as Kneadatite. Do not buy it from "the army painter" or "greenstuff world", they repackage it in a hilariously small pack and raise the price on it a shit load. Apparently you can use "grey stuff" aka procreate as an alternative, but I haven't tried it so I can't vouch for it.

    once you have the greenstuff, you've got options:

    • mix the greenstuff with milliput - I use a 50:50 ratio and its really nice. Greenstuff on its own cures to a quite soft consistency, and you can't sand it. adding milliput makes it cure rock hard. you can just mix all four parts together in one go, you don't have to do 2/2.
    • put on a layer of greenstuff, then before it cures add some polymer clay (sculpy super firm or fimo). Polymer clay doesn't bond to the armature at this scale on its own, but the greenstuff will give you a bond you can work with. After that you can just layer up with poly clay as much as you want
    • work in straight greenstuff - I've done this a little bit, and it has a different feel than working with the milliput mix.
    • some people work in straight milliput, and I honestly don't understand how they get it done. Though I think thats how the oldschool people did their hard surface modeling.
    • you can work in straight poly clay if you are going up in scale, just use foil as your armature, and it will bond fine. Or no armature at all if you are doing little lumpy dudes that don't need it.

    Also, note: the hobby shop sold me some AK brand greenstuff that turned out to be liquid greenstuff. This is not a two part epoxy. You will not get anything done with this. Make sure you don't buy this (unless you know what its for and you are specifically buying this stuff)

    VASELINE / Nivea hand cream The tools will stick, and it sucks asssssss. You need to get some lube on your tools. I've tried using water and spit, but neither of them work great. Also, when working with any miliput, water causes it to get silty and run off your model. I also tried eucerin, because that was the moisturizer I had on hand, and it didn't do anything nice either. Anyway, try out some vaseline, and you will get the nice smooth surface texture you are looking for. If you are using polyclay, you cannot use water, you need to use vaseline, otherwise it will get all crumbly when you bake it.

    tools

    Get yourself some clay shapers, and don't get the shitty ones from amazon. The amazon ones are way way way too soft. The pack I got was the royal sovereign size zero extra firms from blick, but I also had one of the non-extra firms and it was pretty good. Unfortunately these are quite expensive.

    Scalpel / exacto knife - I've been using a scalpel with a number 11 blade. I find it nicer than the xacto, because its much cheaper, and the blade is quite a bit flexier, so you can use it remove stuff from a tile. Also xacto blades break all the time, and once you lose the tip you lose a ton of the value. On the downside I've been sculpting in bed and keep dropping a scalpel onto my chest, sometimes pointy end down. Don't do this.

    A piece of tile for shaping things on - I'm using a tiny little marble sample from a flooring place. before I was using this, I was using parchment paper, but its much nicer to just have something ready to go. If you've got two of them, even better, you can make a consistently flat object that way.

    some cork - find a wine drinker. Make sure you cover the top with a layer of epoxy before working on it, if you attempt to adhear the models feet directly to the cork surface, they will detach, and you will be driven mad by your floppy model.

    something for manipulating the armature wire - a pair of needlenose pliers and a way to cut the wire.

    above are the only critical things I think - you can try out small wax tools / dental tools, but I wasn't having the greatest time with them. Lotsa people fashion their own tools, I made a hooky poker by filing down a piece of my armature wire, bending it, and epoxying it to a old paint brush.

    A ball end pokey thing is real nice for adding texture - blick sells this as well. Its not quite a needle, its got a round tip. Again, get the smallest one you can.

    I have a set of 10g makeup containers that I use for storing all my materials in my little sculpting kit. Not necessary, but cool that this whole thing (including the piece you are working on) can get rolled up in a bag that is about the size of a sketchbook.

    While sculpting in bed, I've been using a super bright camping headlamp. I bet I look really cool.

    sculpting notes

    Its bonkers how tricky it is to work at this scale! While I'm finding that I can work the underlayers a bunch, for the upper layers and detailing, I find that I have to get it right the first go, or scrape the messup off and try it again. I end up working off the model quite a bunch, and transplanting the bits I've made onto the model. So stuff like strips, or cones, or little rolled up pieces cut to the correct size (bit of tile and scalpel for all this cutting).

    There are tons of "tricks" if you dig around - stuff like making chainmail, dragonmail, chains, or skulls... wh40k is the locus :D but even if you aren't doing the tricks, the concepts are repeatable for other outcomes.

    The vaseline for smoothing a blending is critical. I kept getting horrible surfaces and wondering why, and forgetting that I needed to grease up my tools. Apparently it can make it difficult to bond the layer after the layer that you smoothed with vaseline, but thats a problem for a different day imo. Folks told me to use water in the final pass? I dunno, if you figure out the answer here let me know, I'm still experimenting.

    working in multiple passes is the way to go - you can speed up cure time (generally a few hours) by putting the piece under an incandescent lamp. I'm using a costco peanut can with a clamp lamp balanced precariously over it - this drops the cure time down to ~10 mins? Maybe less? I'm not sure, I actually keep turning my pieces brown with this assembly, but they don't seem to have gained any brittleness. This is not critical, you can also just chill or work on multiple models.

    I got some models that I know are scupted by hand with these techniques so that I could have a scale reference - the perry miniatures. Probably better to just hit up a pal that has minis on hand (thx kyle for showing me). Modern minis are nuts though, due to computer modeling. Its not just scale ref, the pros have mind melting technique, its cool to see the upper limit.

    There are tons of ways to make armatures - I've been doing mine all in one (arms and legs together), but if you have a pin vise (a tiny drill) and super glue, you can do the main body first, then add in the arms. I can see the argument for this, its kinda annoying to work the body while you have these arm bits flopping around.

    casting

    It cost me about 200 dollars to get all the bits I needed for casting in pewter. I suspect I could have cut some corners and done it a bit cheaper, but I didn't want to be fiddling around with a stainless ladle and a propane hand torch. I followed the recipe in this video: https://www.youtube.com/watch?v=h6kh2RJl1dQ

    The parts that are difficult to compress in the price are the high temp mold material, and PPE. You can theoretically melt down pewter you have on hand, but good luck lead testing it all. Its possible that its cheaper to work with resin, but I'm not interested to find out.

    some loose thoughts

    • the molds don't work until they heat up, so run a few casts through them before getting disheartened. you can chuck your bad stuff back in the pot and melt it down again. It may be possible to do this in an oven, but I don't wanna mix food and what the mold is made out of.
    • use baby powder in the mold to get stuff into the nooks and crannies
    • the weight of the metal in the pour sprue is critical to filling the mold. Make sure it is wide enough.
    • the main pour passage needs to be wide enough to fill out the rest of the mold as well. don't make the mistake I made of giving your figures spindily legs.
    • working to the mold line is kinda fun! at it immediately makes your models look like old school warhammer.
    • I didn't have any degassing chamber for getting the bubbles out of my mold material before pouring. Turned out fine somehow?
    • cast minis feel fuckin awesome. Like a big chunky worry stone. Hell yeah.

    safety

    I didn't dig too much into the MSDS for all the stuff I'm working with, but here is my general understanding:

    • contact dermatitis is a risk with these epoxies. I haven't got it yet, but some people do. I'm not wearing gloves when working with it so far.
    • many of the putties, when cured, are food safe. Again, I'm not eating off of them, but they are marketed for both sculpting and plumbing work. So that end isn't that scary.
    • they do have carcinogens in them, in particular: talc. I'm not too worried about this, because its contained within the putty when working, and when sanding it you can ppe up and keep the dust contained. Its also much lower quantities percentage wise than baby powder and quartz countertops, which seem like the main ways of getting cancer from talc.
    • its plastic in the end, so if you are sanding, microplastic explosion. Wear PPE.
    • if you go down the casting road: WEAR FUCKIN GOGGLES + gloves. pewter will spit if it gets into water or gets water into it.
    • I'm also kinda grossed out by the rubber molds I'm using, they've been offgassing for weeks - no one mentioned this, and I find it pretty gross. Its not much, but enough.

    so far, I'm just washing my hands before and after, and not putting my tools in my mouth to wet them (which is what some people recommend... classic way to injest poison imo). You can dig up MSDS for all this stuff, and its all been around for 40+ years.

    loose resources


  • Jonathan Brodsky - Fri 11 November 2022 - projectlog, synthesizer

    I made a quick video to show off a bit more about this synth for folks that are into visuals / audio for learning.

    Also, I got reverb working after I made this video, so here is a tune with reverb

    https://www.instagram.com/p/Ck39LkPsGd0/


  • Jonathan Brodsky - Tue 01 November 2022 - projectlog, synthesizer

    Since my last post ... many things have happened:

    I shared a prototype with a friend and got some good feedback. This was a huge milestone for me.

    I discovered that I really don't like the buttons! So many indie electronics projects use them, but I found that they could tear off quite easily, with just a little bit of lateral force. I did a test board to try out ninja flex clear buttons - this feels like close to the correct solution - I love the brightness, and the feeling is pretty good. And most importantly, they feel much more solid than the existing buttons.

    The downside of course is that its more cost per unit, but I'm ok with that trade off. The second downside is slightly less front panel lableling space, but that's probably ok.

    I've mostly done work on the firmware since the last update. I feel like I'm closing in on a feature set now:

    • 8 channels of sampler / synthesizer (organized in two banks that mute their pair when triggered) or 16 midi. These assignments can be set per channel.
    • per pattern / channel trigger rate and loop length, 1-64 step loops, 2x-1/8x trigger rate, with a smattering of triplet rates in there
    • ~3 minutes of sampling capacity (I might make this lower to add a "song mode")
    • midi in / out sync
    • channel muting mode (like your old roland groovebox)
    • per step parameter locks (2048 per song, so you can run out ... but I'm running out of ram here)
    • per channel effects send - I've got a basic delay in, I think I can afford some other cheap effect, like a chorus or LUT based distortion

    The above is more or less working. Below are things that I'd really like to finish before the next hardware rev

    • 2 assignable AD envelopes, 2 assignable LFOs per channel - I've prototyped this, but I need to put in the ui for the assignments and make everything work more predictably.
    • midi note in - I've tested this, but there are some bad decisions with how I store notes that I need to unwind.
    • slides / legato for both internal synths and midi
    • PO / volca click sync support

    I'm also planning on cutting some features, in particular the onboard speaker / amp. I don't feel like it really makes the unit that much better, and folks will either have headphones or some other speaker. If they want a PO quality speaker, they can buy a PO, they are cheap.

    I'm 90% sure that I've figured out the horrible noise that I'm getting from the addressable LEDs, its just a matter of throwing a little RC filter on the power for them, and it goes away. Unfortunately this also means a voltage drop, but in testing I've found that I can lower the noise a sufficient amount and still keep the lights on. I'm a bit worried that when I put this on the actual unit, it won't actually work, but we will see.

    In October 22, I spent some time making loops with the unit, just to play around with it some. Not all the loops on this page use it, listen to the ones labeled TDM. Try out loop 1 (for a silly sampler workout), 7, and 16.


  • Jonathan Brodsky - Fri 14 January 2022 - projectlog, synthesizer

    Low effort project log. New enclosure, more streamlined assembly (rp2040 direct onboard). Pocket Operator for scale. Been working on firmware a bunch as well.


  • Jonathan Brodsky - Mon 20 September 2021 - projectlog, synthesizer

    This is a short update - projects not dead. Rev 5 has been built and is functional. Quick video here: https://www.instagram.com/p/CRMyKk9BBYB/

    ~~I made the git for the project public, and uploaded schematics for rev 6.~~

    update 11/2/2022 I've now made the project on git not public. The reasons are: that I don't want people ordering these boards with all the issues, I'd like to sell a few of them first before people start ordering them, and I'm unsure exactly the level of opensource I want to be on.


  • Jonathan Brodsky - Thu 30 April 2020 - projectlog, synthesizer

    Assembled rev 3 over the course of the past three days - this one makes some pretty major changes:

    • dropped the AA batteries in favor of a lipo & charge circuit
    • samd10 on the frontboard to manage the buttons & LEDs. this is now fast enough to support PWMing the LEDs
    • fixed the issues with the speaker & amp, actually sounds nice now
    • sd card on sdio (yay, fast, 4bits at a time)
    • line in detection works, headphone still isn't (I swear I worked this out on the breadboard)
    • I put the wrong footprints in for the mic, so that won't work. I might be able to fix if I had hot air rework, but I don't.

    This time I bit the bullet and actually soldered stuff in, rather than putting female headers everywhere so I could rescue the expensive bits. This let me mostly assemble it, with two execeptions. I stacked up the two ffc connectors that are underneath the teensy, rather than offsetting them. Also the ffc connector doesn't actually fit underneath the battery, so I can't fit it inside the sandwich.

    I think both of these issues are fixable.

    The larger question I'm up against is if I want to take a hard turn away from teensy, and just make this an stm32 board. It would majorly lower the complexity in some ways, and definitely lower the cost. I could potentially also make it a single board, rather than this sandwich stackup.

    I think in need to grind on that question for a few more days, while I'm working on the firmware. Luckily the work that I'm doing on the firmware is transferable to whatever arm chip I end up slotting into the thing, so that isn't a huge issue.


  • Jonathan Brodsky - Sat 25 April 2020 - projectlog, synthesizer

    The Basics
    This project started because I thought it would be interesting to build my own pocket operator. There are a few other projects diy projects that come do this as well, the Teensy Beats Shield, the Wee Noise Maker, and the polaron in particular are pretty nice!

    The core was going to be a teensy 4, since it is wildly overpowered as an embedded audio platform, and I expect I can get a bunch of synth channels running at once. Outside of that, the goal is to make everything battery powered, pocket sized, and with a similar ui layout to the pocket operator.

    Rev 1The initial revision had the following modules:

    • A boost chip to get the AA batteries up to 3.3v - I followed the boost section from the powerboost1000.
    • A pam class d amp for the onboard speaker
    • flash memory chip
    • sgtl5000 codec (the same as the teensy audio shield)
    • audio in / out over 1/8 inch jacks

    I also build an io board that had

    • 25 buttons
    • 25 leds
    • 2 encoders
    • and a spot to drop in the oled 96x64 oled display I was using.

    This io board just broke all the various parts out to a giant pin header, and I ran the matrix from an arduino pro mini clone I had lying around. The arduino pro mini was communicated with over i2c from the teensy 4. This was the part of the project that actually worked! I got nice pwm control of the LEDs, I could get fast button responses, all sorts of nice stuff.

    The part that didn't work so well was the back board. I got the boost chip working, and the amp seemed to work (when I fed it external audio) but I irrecoverably messed up the power supply for the codec, so I couldn't get it dumping audio out. I tried changing the configuration of the codec (its a big blob of i2c nonsense you send it) but there was no audio no matter what I did. Turns out the issue was elsewhere.

    You may notice this revision is dusty as heck. The front board was ordered in september, and the backboard was from november.

    Rev 2

    I didn't make a ton of changes to the main board on this one, other than fixing the issue with the power supply for the audio codec. Once I had that up and running though, I discovered that the audio still wasn't working. I went digging, and found that the in and out lines for the i2s interface were flipped. I had copied the schematic from the teensy audio shield, and those had them flipped. Luckily I was able to easily bodge a fix for this.

    you will notice the front board is listed as r.1, I started numbering them at zero, but then changed it. Bad version control, whoops.

    The front board went through more changes. I decided not to bring the pro mini onto the board, but rather use an i2c gpio expander. I didn't do the calculations up front, but it turned out that I couldn't push data to it fast enough to accomplish pwm brightness control of the LEDs.

    I messed up the pin header for the oled, so I couldn't actually solder it into place, and due to the fact that I was adding female headers to socket the teensy, that meant it was too tall for me to mechanically assemble the sandwich anyways. I also ordered the wrong amp chip, it was a class a/b instead of a class d. I don't know that this really affected the output a ton, but I couldn't get a clean sound out of my speaker.

    The good news was that the buttons and lights basically worked, the oled basically worked, and I could write software for it. I started building a sampling drum machine for it, but got kinda bored of writing dsp code I knew had been written a million times before. A few months passed while I worked on other projects, and then I came back to it and decided to port mutable instruments plaits so I could quickly get a ton of really nice sounds.

    I also started thinking about what rev 3 for the hardware would look like. I have the hardware on hand now, and will post another log after I finish doing bringup.