July 9, 2019 at 18:54

To the question of different or another crying of Yaroslavna

    (Chinese) A chicken pecking a grain of it happens sometimes.


    To begin with, I would like to complete the discussion about the cost of production of modules, begun by the previous post. In the comments to it, a reference was given to a module similar to the one under consideration, with the striking difference that it was sold on Ali for an absolutely incredible amount of 58 rubles (I will not use the monetary unit later, the original probably refers to RMB or US dollars, but in relation to the inhabitants of 1/8 of the land, the site kindly conducts conversion). Given the fact that in my post I determined the module’s prime cost at 300+ rubles, the indicated price caused a slight rush (“and this is not a figure of speech, indicating some obscure garbage, but a very specific verb” - well, I liked this expression of Divov very much).


    PNP: KDPV to the second part of the crying.

    However, there is an objective reality given to us in sensations, and this module at this particular price is part of this reality, therefore we will assume that sensations do not deceive us and we will understand. In the module description, again kindly translated into Russian, the TL microcircuit is mentioned, on the basis of which the module is made and in the datasheet (hereinafter referred to as the date) there is an example of a typical implementation on this microchip, which does not contradict the module’s photo in any way.

    This is a specialized mxx for controlling an LED matrix, which contains:

    1. interface part for I2C with address programmable legs
    2. built-in generator and control circuit,
    3. up to 8 keys of the lower shoulder,
    4. up to 12 upper arm current generators
    5. with setting the nominal current value by an external resistor and
    6. individual adjustment of current for each matrix element.

    I'll take a little rest


    The only question that can be asked for the developers of msx is: why line terminators are not included in the msx, it seems to me that the implementation of a resistor and capacitor of these values ​​should not be difficult. Yes, the terminator is needed only at the end of the line, but the additional leg that controls the connection of this solves this problem. And the presence directly on the board of additional elements for coordinating the transmission line, on the contrary, creates a problem, since the external circuit necessary to turn them off will entail excessive costs and is completely absent in the implementation under consideration. Moreover, an additional leg could choose one of two possible addresses. In principle, one can imagine a certain cascading scheme of modules (or msx),

    We have finished circuitry, we are moving on to economics and here everything is terribly simple and obvious:

    1. for definiteness, we assume that the cost of installation in China is 1/2 of the corresponding indicator in the Russian Federation (indicated after +), although this is not so important,
    2. an indicator similar to CA56, on Ali costs 30 + 6 (against 130 in the distributor company from my last post + 12),
    3. the mentioned msx 6 + 5 (60 + 10),
    4. she just does not need current-limiting resistors 0 (8),
    5. 4 matching elements cost 2 (0),
    6. two capacitors have not gone anywhere 1 (2),
    7. the cost of the fee is determined at 12 (10 fees for $ 2) versus 28.

    Total 36 + 11 + 2 + 1 + 12 = 62 - a little more than the selling price for Ali, but pretty close, accepting our assumptions and, which is significantly, five times less than 142 + 70 + 8 + 2 + 28 = 250. That is, being a Chinese entrepreneur, and purchasing components in the neighborhood, it is quite possible to sell modules at a price 10+ times lower than the domestic manufacturer does. The only inconvenience for the buyer is a delivery time of 2+ weeks, which may be a determining factor in favor of import substitution for an electronics lover.

    Crying begins


    The question remains why our distributor costs so much, that is, 5-10 times more expensive, and I know very well that nothing of this module mentioned in the BOM (except for resistors and capacitors) is produced in the Russian Federation and everything is purchased in the same PRC at prices not exceeding the price of Ali. But, as I wrote earlier, I do not have a company of such a scale as Platan, and I have no right to give advice and recommendations on pricing, I just express my bewilderment.

    Well, the second part of the perplexity lies in the price of the components - why we cannot produce them at such a price - and I do not need to talk about cheap labor - the living labor is vanishingly small in the cost of the crystal, the main contribution comes from the depreciation of the equipment, and I don’t think that China gets it at low prices.

    Pnp: It is precisely by misunderstanding and perplexity that I differ from the State Duma deputies and members of the Government of the Russian Federation who do not doubt for a second (they do not doubt absolutely completely, but “this is my personal value judgment, expressed in a private conversation”) that they understand absolutely everything and can set the maximum prices and allowances in various sectors of the economy. Indirect proof of the correctness of my value judgment is the brilliant position in which the Russian economy in general and electronics in particular are under such brilliant leadership.

    The sun used to shine brighter and ...


    Part two of the discussion of everything will be devoted to one interesting crystal (I call microcircuits) manufactured by TI. In Jack Gansley’s blog, I respect him a lot, there is a “product of the month” section in which he reviews interesting new items, and I decided to follow his example. It is not at all necessary that these will be new products, I will simply describe them as they appear in my field of vision.

    Today's product is Seensor Tag 2, as always, from TI. Given the scale of acquisitions by this company of other manufacturers, soon all types of axes will be included in its portfolio. On the one hand, I understand the threats arising from the establishment of a monopoly, on the other hand, I cannot but recognize the conveniences created by the monopoly. In any case, no one was interested in my opinion before the merger of TI and the NS, and therefore I categorically refuse to be responsible for any negative consequences of this event and do not pretend to have positive consequences for laurels, but simply inform about those and others.

    The module itself is very interesting, I ordered a couple and after receiving them I will give a detailed analysis, because I prefer to rely on my own impressions and test results, rather than on press releases from the marketing department (I know their habit of not negotiating, and on an alarming scale). In the meantime, I looked at the module diagram, kindly presented in the technical documentation section and drew attention to one of the components installed on it - TPS22910, then you can follow the date at www.ti.com/lit/ds/symlink/tps22913.pdf

    Briefly about the chip


    This is just a key for a very limited current of 2 A and a voltage of not more than 5.5 V - it would seem what to talk about, take any field worker and enjoy. But the devil, as you know, is in the details and here they are:

    1. the minimum switching voltage is 1.5V with a passing resistance of 80 mOhm - not every field-effect transistor will show such results,
    2. “Logical” and invariable threshold for switching control at any switched voltage,
    3. built-in reverse current protection,
    4. integrated undervoltage protection,
    5. guaranteed (and not bad) temporary switching parameters,
    6. small-sized case (to me, in general, according to FIG, but the manufacturer emphasizes this fact).

    I hope I praised this device enough so that subsequent statements would not seem like a run-in (and they will follow, which of my readers would doubt).

    Pnp: It would be interesting if there were any shortcomings, I would write about this msx - this is a rather rhetorical question, but it makes me think about my motives for creativity. When another interesting crystal with perfectly written documentation falls into my hands, then much will become clear.

    But I did not understand this


    There are only four legs for the microcircuit - input, output, ground and control - which is much easier. So, I look at the circuit board, on the switch-on leg there is a pull-up resistor to the ground, it’s very obvious that it turns on the mxx, I open the date to make sure that this is really so, and I observe an absolutely amazing picture. A step is depicted near the control foot in section 4 (recently such a style has become popular, although it seems to me personally that the truth table is more informative and does not provide space for interpretation) and on this step they usually write on / off (dis / en), but in in this particular case it is written LO / HI. Thank you, captain, but I did not understand what level connects the input to the output.

    The functional diagram in section 10.2 did not clarify, since the control foot is included in the box with the name Control Logic, and it can do anything. Section 10.3.1 once again tells us that this leg controls the state of the key and again does not say a word how.

    I scroll down the date further - in the parameter table (section 8.3) for the control signal the value is indicated as low level (0.6V) and high (1.1V) - it is really possible to control from MK with any reasonable power supply. In addition, section 8.5 indicates the current (1 μA) required to form the required level on the control foot - which is quite suitable for MK with any reasonably designed pins. True, for some reason, the current is given with only one sign (probably, the absolute value is meant) and for voltage values ​​above 1.5V or exactly 0V (that is, nothing is guaranteed at operating values ​​- it's funny). A little further, in the description of the work (section 11.1.2), it is said about the need to maintain the level and inadmissibility of a “floating” input, but again not a word about the inclusion level.

    Further to the date - in section 9, there is a drawing of the methodology for measuring the on / off delay, equipped with the same picture with a step (and the same signatures) and it clearly follows from it that the output is turned on at a high level on the control. An indirect confirmation of this hypothesis is that the name of the control input is given without underlining, underlining, as well as ~ or #, which it is customary to supply inverse inputs. But what about the scheme where the msx is turned on at a clearly low level - we continue the investigation.

    At the very end of the date (section 11.2.3.1) there are oscillograms that give an idea of ​​the dynamic parameters of msx (they give an idea, in no way can they be guided as a data source, for this there are tables with an explicit indication of the parameters) and it can be seen from them that the resolution level is still low. But why is this not explicitly indicated anywhere - we look at the date further and find another set of waveforms (section 11.2.3.2) in which the msx is turned on at a high level. And here comes the clue to the quest - in the family there are four types of msx (22910A, 22912C, 22913B, 22193C), which differ not only in load capacity, but also in the level of inclusion, and the date is common for everyone. For option A, the active level is low, and for the remaining three - high.

    That was the thing


    It turns out that on page 3 the date in section 6 is a table of differences between the performances, in the last column of which this feature is clearly indicated, but who carefully looks at the contents of such tables - so I did not look. We leave out the reason for the decision to make the inclusion levels of different options different; I don’t presume to guess it, probably it should be. But we must pay tribute to the developers of the documentation - they honestly reported this feature of the msx, I just turned out to be not careful enough. Moreover, the heading of the table with the parameters in section 8.5 date clearly shows the on and off levels for the various msx options, which I also did not pay enough attention to (I used to look at the notes at the bottom of the parameter tables for a long time, but missed the heading).

    PNP: I immediately remembered the MCX 1401UD2, which had a power layout of -4 + 11 in a plastic case (K), and 4-11 in a metal-ceramic (KR), which was a particular joy to the developers. Fortunately, the layout of the remaining legs made it easy to deploy the case (type DIP16) and get a working board. So not only foreign manufacturers can successfully make a joke (fun to poke fun).

    Sarcasm Plate


    Well, after I had enough dusted my head with ash and admitted the sin of negligence, I allow myself to predict that if in the picture in section 4 there would be written “see Table 6” next to the step or there would have been a link to a similar note, then no there were no complaints against the developers in principle (probably ... "I think so").

    PNP: I looked at a date from several manufacturers and found discrete P-channel transistors with close parameters (resistance is not more than 88 mOhm with a gate voltage of -1.5 V). It is interesting how they achieve this, my knowledge remaining from the course "Technology of Production of Electronic Devices" to answer this question is clearly not enough, but it would be interesting ...
    Tags:

    Also popular now: