What the seller forgot to mention #5 – the PUPOPAN fish finder, where the bottom is a picture from memory

This is part five of the series in which I check how much of what you see on a fish finder screen is a measurement and how much is a drawing. In the previous episode the Erchang turned out to be gear that measures honestly and draws a real echo, except that switching the fish icons on wipes the entire water column off the image.

Today PUPOPAN, sold as a “New Wireless Fish Finder” and “3.5-inch high precision”. This time no phone and no app: the set comes with its own display in a rubberised case with buttons, a transmitter with an antenna and a transducer on a cable. It looks the most serious of the whole line-up I have covered so far, and that is exactly why it landed on the bench.

The result is the worst in the series and this time it is not about settings or an app. The bottom image on this screen is entirely a drawing baked into memory, while ten bytes come out of the water, two of which are the checksum. Below I show that drawing, reconstructed straight from the display memory.

How this set works

The transducer on a cable hangs under the boat or under a float, the transmitter with its antenna lies next to it, and everything it measures goes by radio to the display in your hand. The display has a 320 x 480 panel, of which 300 by 470 pixels are given to the sonar area, and firmware marked XF08M version 3.2 dated 25 March 2026.

The whole case rests on one question: what exactly does the transmitter send to the display. Because that is all the display has to draw anything from.

Ten bytes come out of the water and not a single echo sample

The frame always has the same length, 10 bytes, in both directions. Here it is in full:

• bytes 0-1: depth
• byte 2: bottom hardness, raw echo strength 0-255
• bytes 3-4: temperature
• byte 5: transmitter battery level and fish size, four bits each
• bytes 6-7: fish depth
• bytes 8-9: checksum

Eight bytes of content. Not a single echo sample. For comparison: the Erchang from the previous episode sends 120 amplitude measurements per ping, six times per second. Here you get one depth number, one hardness number and one fish.

Every decision is made in the transmitter. The display receives a finished verdict and has no way at all to verify it, because it gets no evidence, only a result. Fish size has four states and bottom hardness is quantised into four levels at thresholds 70, 108 and 146.

The bottom on the screen is 21,600 bytes of a ready picture, scrolled round and round

If there is no echo in the frame, where does the colourful bottom image with its texture come from? The display builds each column of the image by one very simple rule:

for each of the 300 pixels of the column:
    if the pixel is ABOVE the bottom  -> write the background colour
    if the pixel is BELOW the bottom  -> write the next pixel from the ready texture

texture phase = phase of the previous column + 1

And that is the whole mechanism. That texture is 21,600 bytes of a ready picture, arranged in 360 rows of 60 bytes, each byte a colour number from a 64-colour palette. Every next column of the image takes a window from that strip shifted by 60 bytes, which is why the bottom “undulates” and looks like structure.

So that there is no doubt, this is what a full image assembled by the same rule looks like: the bottom line from the baked-in table of 358 numbers, and everything below it from that one texture.

The consequence is brutal and worth saying plainly: on this screen you will never see a fish arch, weed or suspended matter. Not because your settings are wrong. The entire water column above the bottom gets a single background colour, because there is nothing in the frame to draw it from. The only thing that appears in the water column is the fish icon.

The bottom hardness from the frame is used to choose a set of pictures. So hard bottom means different graphics, not a different echo.

The fish icon is a sticker drawn from two numbers

About a fish the display receives exactly two numbers: size in four states and its depth. There is no echo strength, no shape, no duration of the return, not even information about whether it is the same fish as a second ago.

A picture is made from those two numbers. The display holds three ready fish drawings, one per size. Once a fish is reported, the display converts its depth into a pixel and paints the drawing column by column, one vertical bar one pixel wide and fourteen high, until it uses up 39 columns. Next to it, if you enable it in the menu, it sticks a bar with a number, that is the depth written beside the fish.

What looks on the screen like a target with some shape is a bitmap from memory. The icon lives for 39 columns, which at fast refresh is over a second and at slow almost five, and during that time it travels across the screen regardless of what is happening in the water.

The image scrolls faster than the data arrives

Inside the display a timer runs a hundred times per second. A new image line is created every few of its ticks, depending on the “Refresh Speed” setting: from 8 lines per second at the slowest to 33 at the fastest.

And frames from the water arrive at most seven times per second, realistically five to six. Drawing is driven by the timer and has no connection whatsoever to when a frame arrived, and when the new depth is zero the column takes the value from the previous one. So the density of the image is the density of drawing, not the density of measurement. Smooth scrolling suggests continuous measurement, and the measurement is not continuous.

After eight seconds of radio silence the display clears the readings and repeats its query. In shallow water sensible readings start only from about 0.6 metres, there is nothing to look for above that.

The radio is better than it needs to be, and that is the whole irony

The transmitter and the display talk over LoRa on 434.000 MHz, with spreading factor SF9, a 250 kHz bandwidth and 4/5 coding. That is a proper radio, not the primitive transmitter from the sondes of the first episodes.

A ten-byte frame takes 72 milliseconds on air. The display answers every frame with its own, also ten bytes, so a full exchange takes 145 ms. That gives a ceiling of about seven exchanges per second and that is the only reason the image cannot be faster.

And now the irony. With a faster setting of the same radio, 240 bytes of raw echo samples would take about 55 milliseconds, that is less than ten bytes take today. A real echogram would fit through this link with room to spare, at five to ten images per second. It is not the radio that is the limit, it is a design decision.

• There is no address or identifier in the frame. Every unit works on the same frequency with the same configuration, and the display accepts any frame with a correct checksum. Two sets within range will interfere with each other.
• The transmit power set in this kit is 100 mW. In Europe this band allows 10 mW of radiated power. How much actually leaves the antenna depends on matching, but the setting alone is ten times over the limit.

The spec sheet against what is actually inside

The promise in the listing	How it really is
“DUAL-BEAM SONAR”	two lines lower in the same table: “Frequency 125 kHz, Single Freq”. The listing contradicts itself on a single graphic
Beam Angle 105 degrees	at 125 kHz the wave in water is 12 mm, and such an angle would require a transducer of about 8 mm, smaller than its own wave. This is not a real beam angle
Max Depth 50 m / 160 ft	correct. The range ladder ends exactly at 160 feet, and readings above 180 feet are rejected
400 m radio range	plausible for LoRa at this power setting, and it is the strongest side of the kit
“Accurate Sonar & Fish Alarm”	the alarm does exist, triggered by the same number that draws the icon

Verdict: one star out of five

What I rate	Rating
Radio: LoRa 434 MHz, real range and resistance to interference	★★★★☆	4 / 5
Depth: above 0.6 metres measured properly	★★★☆☆	3 / 5
Bottom hardness: it is in the frame, but it only picks a picture	★☆☆☆☆	1 / 5
Bottom image: a texture from memory, not an echo	★☆☆☆☆	1 / 5
Water column image: a plain background, no arch and no weed	☆☆☆☆☆	0 / 5
Is it worth buying as a fish finder	★☆☆☆☆	1 / 5

Check it yourself, on your own water

1. Anchor the boat over a flat bottom and watch the screen for a minute. The depth stays put while the texture of the bottom keeps flowing. That is a scrolling texture, not a change in the bottom.
2. Change “Refresh Speed” from the slowest to the fastest. The image scrolls four times faster and not one bit of information is added.
3. Go over submerged weed. You will not see it. Ever.
4. Switch on “Simulator”. If you cannot tell it apart from work on the water, that is exactly the message.
5. Walk away with the display until you are out of range. After eight seconds the readings reset to zero.

Summary – what PUPOPAN does not say out loud

• There is not a single echo sample in the frame, so there is nothing to build an echogram from.
• The bottom on the screen is a scrolling texture of 21,600 bytes, the same for every water and every lake.
• The whole water column is a plain background. You will never see weed, a thermocline or a fish arch.
• The fish icon is a ready drawing in three sizes, painted from two numbers across 39 screen columns.
• Bottom hardness picks a picture, it does not change the echo image.
• The image scrolls several times faster than the data arrives.
• Demo mode looks identical to work on the water, because it uses the same drawing routine.
• No addressing and no encryption, and a power setting ten times over the European limit for this band.

And the other side of the coin, because credit is due: the radio here is genuinely good, depth above 0.6 metres is measured properly, the transmitter battery level is reported, and the graphics layer, that is the history buffer, the palette and the hardware blit, is done sensibly. If raw samples went through that link, this would be decent gear. They do not.

As a depth sounder with a fish alarm and its own screen, with no phone and no app, this set makes sense and may be enough for some. As a fish finder, that is for looking at the bottom, weed and structure, it is not worth spending money on, because the image you see does not come from the water.

What I did not test and therefore do not write about: the transmitter and the transducer, because I have no dump from them, the real power at the antenna, the range on the water and how the transmitter makes the “this is a fish” decision. I describe only what I could establish: what comes out of the water and what the display does with it.

Sources and material worth checking

• NOAA – What is sonar? – how a fish finder works: a transmitted pulse, the echo return time, the conversion into distance
• FAO Fisheries Technical Paper 240: Fisheries Acoustics – beam angle, target strength and why the loudness of an echo says nothing about the length of a fish
• Episode 4: Erchang XF07C – a sonde that sends 120 real measurements and draws the echo from them
• Episode 3: the black ball XJ02B
• Episode 2: Lucky FF918

In the next episode: a sonde a class above, which sends 300 echo measurements from the water and not a single fish field: the Vexilar SP200, the first sonde in this series that hands you a measurement.

What equipment I use and on what basis I test this

Every device described in this series was bought with my own money and is my property. I examine them in order to know what really reaches the angler’s hands and what I can demand from the hardware I design myself. The analysis covers one specific unit and the firmware version I found in it; other batches, versions and variants of the model may behave differently.

Observing, studying and testing the functioning of a program in order to determine its principles is permitted for a person who lawfully acquired the device: article 75(2)(2) of the Polish Copyright Act, which implements article 5(3) of Directive 2009/24/EC. Reading the transmission concerned exclusively my own transmitter and my own receiver. I do not publish source code, memory dumps, keys or tools that would allow anyone to copy someone else’s software – only conclusions about what the hardware measures and what it does not.

Names and trademarks belong to their owners and are used solely to identify the hardware under test, within the scope of critical analysis and the right to information. Photographs from sales listings are used for the same purpose. If a manufacturer or a seller believes any conclusion is wrong, I will gladly repeat the measurements on a unit they provide and publish a correction.

That is how it works here: before anything goes into our boats, it passes across this bench. And if something does not pass, we say so out loud, even when nobody wants to hear it.