E-STIM 2B tear down review

The most versatile and user friendly E-Stim control unit available today. If you want More Power, More Control, and more fun, then the 2B is the one you want.
phxxis
Active
Posts: 12
Joined: Fri Sep 28, 2018 7:11 pm
Location: Denmark

Re: E-STIM 2B tear down review

Post by phxxis »

For all it is worth, I think it is wrong to evaluate stim systems solely based on their hardware. Most of what is happening in a digital box is in the software, inside the microchip firmware. And the development time is probably going to be spent on that software too. So the amount of hardware, or its price in raw components, will not be the full story.

IIRC the ET box is based on AVR-family ATmega whereas the the 2B uses a PIC18-family chip. These are, roughly, in the same ballpark of features. My guess is they either use the 10-bit ADC for audio input, or they use the CCP modules to detect a frequency and pulse width of the incoming signal. Output is probably going to be sent to a DAC, or maybe just a PWM+Resistor+Capacitor (this forms a poor-mans DAC). I'm not sure the pulses needs that much fidelity here. It is likely they generate their own output signal rather than replaying the sampled signal they get from the input (though this could be fun to try).

The original poster also lamented the low amount of memory on those controllers (1536 bytes). There is a design reason for this. AVR and PIC18 are "Havard architectures" so they split data memory from program memory in contrast to mid-to-upper level CPUs which have unified program and data memory (von Neumann archs). They usually have something like 32-64 kilobyte program memory, and rarely above 2 kilobyte data memory. But this means you can use SRAM for the data memory which is going to be very fast. Your laptop often have a fairly small SRAM cache hierarchy at 32 kilobyte at L1. And the main memory is DRAM, which is far slower. DRAM is cheaper and packs more memory in the same physical space however.

The reason you want to have a fast SRAM data memory is because the architecture is a accumulator-memory based one, and not a load-store architecture. So a lot of the operations involves memory transfer. This also somewhat limits the clock cycles of the CPU to the 20-40 mhz ballpark.

(Sidenote: you can often use the program memory for lookup tables and such, so they are not taking up space in the data memory. For instance: text strings for the LCD display, output map tables, and so on will reside as part of the program).

Both of these chip families are outright scary. They execute 8bit instructions at 20-30 megahertz and they often have a single-cycle multiplier at 16bit result width. They are much faster than what is needed to get a decent sample rate out of this, especially if you drop to assembler level.

Personally, I much prefer the AVR instruction set over the PIC one. I see what they were trying to do with PIC, but it feels so broken to write at the assembly level. Also the AVR C-compiler tooling is more free, which eventually led to something like the Arduino adopting it I think.


User avatar
KennyT
Active
Posts: 84
Joined: Sun Aug 13, 2017 3:40 pm
Location: Florida, USA
Contact:

Re: E-STIM 2B tear down review

Post by KennyT »

WHY IN THE HELL WOULD YOU SPEND THIS KIND OF CASH ON A HIGH END PRODUCT WITH LIFETIME WARRANTY (THAT MEANS WARRANTY FOREVER) AND VOID IT BY TAKING IT APART?

I am as curious as the next person but DAM. How stupid can you get? Ok maybe stupid is harsh. But i have way to much money invested in this product (that works flawlessly every time I hook it up) to even think about taking it apart.

BUT DUDE IF IT STOPS WORKING????

I send it back! They fix it and send it back to me. DAM!!
END OF STORY!

Who the hell cares what is inside. If it is full of rocks or electric eels, Who the hell cares, as long as it works like this!

I'M GOOD WITH ROCKS OR EELS !!!

:lol:
KennyT
Post Reply