A real parameter table, which is rare here
Most listings on this site leave the parameter block empty. This one has seventeen rows, and they describe the product rather than the freight. That deserves to be said first, because the rest of this article is about the disagreements, and the disagreements only exist because somebody bothered to publish numbers at all.
The probe is a wireless 3D touch probe for a machining centre: an 8 mm shank, a 42 mm body, a 3 mm ruby ball on a stylus, 2.4 GHz radio to a separate receiver that wires into the control's digital inputs. The model is GP-800.
The seventeen rows, as published
One row is mistranslated: Wood quality is the material field, and the value 304+7075 means 304 stainless steel with 7075 aluminium. Nothing about wood is involved. Every other row reads correctly.
| Parameter | Value as published | What it constrains |
|---|---|---|
| Wood quality | 304+7075 | Material: 304 stainless plus 7075 aluminium |
| Outside diameter | 42 mm | Tool-change clearance and body stiffness |
| Probe thread | M3 mm | Stylus replacement |
| Probe length | 30 mm | Reach and deflection |
| Probe accuracy | 0.001 mm | See the three-level section |
| Probe ball | 3 mm | Feature access and probing force |
| Clamping diameter | 8 mm | Which collet or holder you need |
| Clamping length | 25 mm | How much shank is gripped |
| repeatability | 0.001 mm | The figure that repeats on the same touch |
| XY stroke | 15 deg | Defence against a crash in X and Y |
| z-stroke | 2.5 mm | Defence against a crash along Z |
| Off-line work | yes | Whether it works with the spindle stopped |
| Wireless operation | yes | No cable through the spindle |
| Connection mode | Wireless connection | Radio, not optical |
| Waterproof class | IP67 | Coolant exposure |
| Receiver voltage | DC 5-24v | Which supply you can tap |
Six places where the drawings disagree with the table
The listing carries eight images. Five of them carry technical content: a hero image with headline claims, an installation drawing, the parameter table, a zero-setting drawing, a dimension drawing, a wiring diagram and an application drawing. Read together, they contradict each other six times.
| Item | One drawing says | Another drawing says |
|---|---|---|
| Waterproof class | IP56, hero image | IP67, parameter table |
| Probe thread | M3, parameter table | M4, dimension drawing |
| Probe length | 30 mm, parameter table | 36 mm, dimension drawing |
| XY stroke | 15 deg, parameter table | +/-20 deg, application drawing |
| Z stroke | 2.5 mm, parameter table | +/-3 mm, application drawing |
| Accuracy | 0.001 mm, parameter table | 0.01 with no unit, application drawing |
None of these is a rounding difference. IP56 and IP67 describe different tests. M3 and M4 are different threads. Thirty and thirty-six millimetres are different styli. Fifteen degrees and plus or minus twenty degrees are different crash strokes. Two and a half and three millimetres are different Z strokes. One micron and ten microns are different claims.
A buyer cannot resolve any of this from the page. All six have to be settled in writing before an order, and each has a different consequence.
One micron is what the body says
The probe body itself is laser-marked Repeat accuracy 0.001mm. That is the one figure printed on the hardware rather than added to a marketing graphic, and it is a plausible figure for a probe in this class: one micron of repeatability is what a touch probe of industrial grade is expected to hold when it touches the same point the same way.
Repeatability is a narrow claim. It says that if you touch the same spot ten times from the same direction at the same speed, the trigger point will not vary by more than a micron. It says nothing about whether the micron is in the right place.
Five microns is what the setup drawing demands
The PROBE ZERO drawing explains how to true the stylus: rotate the spindle, adjust four hex screws at 1 N m each, until the concentricity of the ball head is under 0.005 mm, checked against a dial indicator reading in ten-micron divisions.
That is five times the repeatability figure, and it is a manual adjustment performed with four screws. It is also the step that decides whether the repeatable micron lands on the machine's centre line or five microns away from it. A probe that repeats perfectly but sits off-centre produces a perfectly repeatable error.
Ten microns is what the application drawing admits
The application drawing states reciprocating positioning accuracy 0.01, with no unit. Read as millimetres that is ten microns, and it is the only figure on the page that describes what the installed system does rather than what the probe does internally.
The progression is worth holding in mind as a single picture.
| Level | Figure | In microns | Relative to the smallest | Where it comes from |
|---|---|---|---|---|
| Probe repeat accuracy | 0.001 mm | 1.0 um | 1.0x | Printed on the probe body |
| Zero-setting concentricity | 0.005 mm | 5.0 um | 5.0x | The PROBE ZERO drawing |
| Reciprocating positioning | 0.010 mm | 10.0 um | 10.0x | The application drawing |
The span is ten to one. A shop that buys on the 0.001 mm headline and then measures a part to 0.01 mm will conclude the probe is faulty. It is not. It is behaving exactly as the third drawing says it will.
Which number applies to your part
For most work the answer is: none of them alone. The tolerance you can hold is the sum of the probe's repeatability, the zero-setting error, the machine's own positioning error and the thermal state of the part. The probe contributes one micron of that sum directly and up to five microns through its setup.
Where the probe dominates is in comparative work: finding the centre of a bore, setting a datum, checking that a feature has moved between two setups. There the systematic errors cancel and repeatability is what you get. Where the probe does not dominate is absolute measurement of a feature against a tolerance of five microns, on a machine that has not been thermally settled.
This is not a criticism of the GP-800. It is the same arithmetic that applies to any touch probe, including the ones whose names appear in this listing's title.
IP56 or IP67, and why the difference is not cosmetic
The hero image says waterproof IP56. The parameter table says IP67. The two are different tests, and the gap matters on a machining centre.
IP56 is protection against dust in quantity sufficient to interfere, plus water projected in strong jets from any direction. IP67 is protection against dust ingress entirely, plus immersion in water at a stated depth for a stated time. A probe on a machining centre meets coolant as a jet, not as a bath, so IP56 is the specification that actually describes the environment.
IP67 is the stronger claim, and it is the one in the table. If the probe is genuinely IP67 then IP56 is simply an understatement on the graphic, which is harmless. If the truth is IP56, then the table is overstating, which is not. The only way to know is to ask, and the question is cheap.
One more thing about the word: waterproof is not a rating. IP56 is not waterproof, it is jet-resistant. Coolant at pressure will find a seal that a splash will not.
Stroke in degrees, and what that means at the ball
The parameter table gives XY stroke as 15 degrees and Z stroke as 2.5 mm. The application drawing gives XY as plus or minus 20 degrees and Z as plus or minus 3 mm. Degrees for a linear-sounding quantity looks wrong at first, but it is the right way to describe this mechanism: the stylus pivots, so its travel is an angle, and the linear movement at the ball follows from the stylus length.
| Stylus tip travel | At 15 degrees | At 20 degrees |
|---|---|---|
| On the 36 mm stylus | 9.32 mm | 12.31 mm |
| What the parameter table claims | 15 deg | not stated |
| What the application drawing claims | not stated | +/-20 deg |
On a 36 mm stylus, 15 degrees puts the ball through 9.32 mm and 20 degrees through 12.31 mm. Both are plausible crash strokes, and the difference between them is three millimetres of travel before something bends.
Whichever figure is true, the safety stroke is small compared with the travel of a machine that moves at feed rates measured in metres per minute. The stroke protects against a slow approach error, not against a rapid move into a fixture. That is what the break-shaft stylus is for, and the listing sells the stylus as a separate line item resolution.
The 3 m radio limit against machine size
The receiver carries a printed warning: The distance is less than 3m. It also says the side must face the spindle direction, which is a mounting orientation constraint, not just a range constraint.
| Machine envelope | Body diagonal | Share of the 3 m radio limit |
|---|---|---|
| 1.0 x 0.6 x 0.6 m | 1.31 m | 44% |
| 2.0 x 1.0 x 1.0 m | 2.45 m | 82% |
| 3.0 x 1.5 x 1.5 m | 3.67 m | 122%, over the limit |
A small vertical machining centre at one metre by 0.6 by 0.6 has a body diagonal of 1.31 m, so the receiver sits comfortably inside the limit wherever it is mounted. A two-metre machine at 2.45 m diagonal uses 82 percent of the range before any margin for the metal of the enclosure, the coolant and the tooling between probe and receiver.
A three-metre machine exceeds the limit on the diagonal alone. On such a machine the receiver cannot go on the outside of the enclosure and cover the whole working volume; it has to go inside, and then the metal and the coolant are between it and the probe.
Nothing in the listing states a range figure other than that warning. No frequency channel plan, no statement of what happens when two probes run in the same shop, no guidance on receiver placement beyond the orientation note.
Forty-two millimetres across, eighty-nine millimetres out
The dimension drawing gives an 8 mm shank, 25 mm of clamped length, a 53 mm body and a 36 mm stylus. Overall that is 114 mm, with 89 mm unsupported below the clamp.
An 89 mm overhang on a 42 mm body is a stiffness ratio of about 2.1 to one. That is not a complaint about this probe in particular, it is the geometry every probe of this shape shares, and it is why probing feed rates are slow and why the approach direction matters.
The 42 mm body also has to clear the tool changer. An 8 mm shank inside a 42 mm body leaves 17 mm of radial overhang on each side, so the probe occupies a pocket sized for a much larger tool than its shank suggests. Check the pocket, not the taper.
Two metals, two expansion rates
The material row says 304+7075: stainless steel and aluminium, in one body. Their coefficients of thermal expansion differ, 17.3 against 23.6 millionths per kelvin, and a body built from both will develop a changing fit as the machine warms.
| Temperature change | 304 stainless grows | 7075 aluminium grows | Difference | As multiples of 1 um |
|---|---|---|---|---|
| 5 K | 3.63 um | 4.96 um | 1.32 um | 1.32x |
| 10 K | 7.27 um | 9.91 um | 2.65 um | 2.65x |
| 20 K | 14.53 um | 19.82 um | 5.29 um | 5.29x |
At a ten-kelvin swing across the body the differential is 2.65 microns on a 42 mm diameter, which is 2.65 times the repeatability figure the probe is sold on. Whether that reaches the measurement depends on where the joint is and which metal carries the stylus seat, and the listing does not say. The point is not that the probe is bad; it is that a one-micron claim and a two-metal body are in tension, and the resolution is a warm-up routine, not a specification.
The wiring diagram names I/O addresses but not the control
The receiver cable has six cores. Black is 0 V, red is +24 V, brown goes to Y2.0, green to X4.7 labelled SKIP, yellow to X4.5 labelled PROBE LOW BAT, and white splits to X4.3 labelled PROBE ERROR and X4.4, which carries no label at all.
Those addresses are specific. X and Y addresses in that form belong to a particular family of controls, and the listing never says which family. A buyer with a different control has to map the six cores onto whatever inputs are free, and the labels SKIP, LOW BAT and ERROR tell them what the signals mean but not what voltage or polarity they present.
Two practical consequences. First, the probe needs three digital inputs, not one, to use the battery and error reporting; a shop that wires only the trigger loses both warnings. Second, the cable is a six-core unscreened-looking bundle in the photograph, carrying a fast trigger signal next to a spindle drive. Shielding and a single-point earth are worth asking about explicitly.
X4.4 having no label is either a spare or an omission. Ask which.
What the title borrows
The listing title, and the titles of its two sibling listings, include the names Marposs and Renishaw alongside Cyclmotion and GP-800. The product is branded TEBAK and the body carries no other manufacturer's mark.
Those two names belong to established metrology manufacturers. Their presence in a title does not establish that the GP-800 is compatible with their receivers, their styli or their software, and it does not establish that it is a copy of any particular model. What it does establish is that searching for either brand may surface this listing.
For a buyer the practical question is narrow and answerable: does the GP-800 work with the stylus thread and the receiver protocol I already have, or am I buying a closed system? Ask for the stylus thread and the radio protocol in writing. If the answer is M3 and a proprietary 2.4 GHz link, then the brand names in the title are decoration.
Three listings, one product
This record is one of three listings in the same queue that carry the same eight images, pixel for pixel, and the same GP-800 model line. Two of them name Marposs and Renishaw in the title, one does not. They are one product listed three times, not three probes.
For a buyer that is useful rather than confusing: three listings mean three price points for the same article, and comparing them is free. For anyone tracking specifications it means the same seventeen rows and the same six contradictions will appear three times, so settling them once settles them everywhere.
If you are comparing this against a wired trigger probe rather than another wireless one, the CF-38 3D trigger edge finder guide covers that class, including what a wired probe gives up and what it keeps.
What to put in the request for quotation
Ten questions, and the first five exist only because the page contradicts itself.
| Item to confirm in writing | Why it decides the purchase |
|---|---|
| Which accuracy figure the quotation refers to | Three numbers appear: 0.001, 0.005 and 0.01 mm |
| IP56 or IP67 | The hero image and the table disagree; only one can be true |
| M3 or M4 stylus thread | Decides which replacement styli you can buy |
| Stylus length, 30 or 36 mm | Changes tip travel, deflection and safe approach |
| XY and Z stroke, stated once | Two drawings give two different sets |
| Radio range and receiver placement rules | The warning says under 3 m; a 3 m machine exceeds that |
| Battery type, life and whether it is user-replaceable | A white indicator warns of low voltage; nothing states the cell |
| Which CNC control the X4.3/X4.5/X4.7 addresses belong to | The wiring diagram gives addresses but not the control family |
| Trigger force in newtons | Not stated anywhere; it sets the deflection before the signal |
| Whether the Marposs and Renishaw names in the title imply any compatibility | Both marks appear in the listing title; the product is TEBAK |
Frequently asked questions
Is 0.001 mm the accuracy of this probe?
It is the repeat accuracy printed on the body: touch the same point the same way ten times and the trigger repeats within a micron. It is not an absolute accuracy claim over the working volume. The listing also shows 0.005 mm zero-setting concentricity and 0.01 mm reciprocating positioning, which are the setup and system figures.
Why does the page give IP56 in one place and IP67 in another?
The hero image says waterproof IP56, the parameter table says IP67. IP56 is dust plus strong water jets; IP67 is dust-tight plus immersion. On a machining centre the coolant arrives as a jet, so IP56 describes the real environment. One of the two figures is wrong and only the seller can say which.
The XY stroke is given in degrees. Is that a mistake?
No. The stylus pivots, so its travel is angular, and the linear movement follows from stylus length. Fifteen degrees on a 36 mm stylus moves the ball 9.32 mm; twenty degrees moves it 12.31 mm. The table says 15 degrees and the application drawing says plus or minus 20, so ask which.
Will it work on a large machining centre?
Check the receiver warning: distance less than 3 m, with the side facing the spindle. A 2 m machine has a 2.45 m body diagonal, which uses 82 percent of that before any obstruction. A 3 m machine reaches 3.67 m and exceeds it, so the receiver has to go inside the enclosure.
What stylus does it take?
The parameter table says M3 thread, 30 mm long, 3 mm ball. The dimension drawing labels the thread M4 and measures the stylus at 36 mm. Both cannot be right, and the answer decides which replacement styli you can buy.
Does it need a battery, and can I change it?
It is wireless, and the indicator panel includes a white lamp for battery low voltage, so a cell exists inside. The listing never states the cell type, the running time, or whether the user can replace it. Ask, because a sealed cell turns a consumable into a return.
Which CNC control do the X4.3 and X4.5 addresses belong to?
The wiring diagram gives X4.3 for probe error, X4.4 with no label, X4.5 for low battery, X4.7 for skip, and Y2.0 for the trigger, but never names the control family. On a different control you map the six cores to free inputs yourself, and you need to know the signal voltage and polarity to do it.
Frequently asked questions
- Is 0.001 mm the accuracy of this probe?
- It is the repeat accuracy printed on the body: touch the same point the same way ten times and the trigger repeats within a micron. It is not an absolute accuracy claim over the working volume. The listing also shows 0.005 mm zero-setting concentricity and 0.01 mm reciprocating positioning, which are the setup and system figures.
- Why does the page give IP56 in one place and IP67 in another?
- The hero image says waterproof IP56, the parameter table says IP67. IP56 is dust plus strong water jets; IP67 is dust-tight plus immersion. On a machining centre the coolant arrives as a jet, so IP56 describes the real environment. One of the two figures is wrong and only the seller can say which.
- The XY stroke is given in degrees. Is that a mistake?
- No. The stylus pivots, so its travel is angular, and the linear movement follows from stylus length. Fifteen degrees on a 36 mm stylus moves the ball 9.32 mm; twenty degrees moves it 12.31 mm. The table says 15 degrees and the application drawing says plus or minus 20, so ask which.
- Will it work on a large machining centre?
- Check the receiver warning: distance less than 3 m, with the side facing the spindle. A 2 m machine has a 2.45 m body diagonal, which uses 82 percent of that before any obstruction. A 3 m machine reaches 3.67 m and exceeds it, so the receiver has to go inside the enclosure.
- What stylus does it take?
- The parameter table says M3 thread, 30 mm long, 3 mm ball. The dimension drawing labels the thread M4 and measures the stylus at 36 mm. Both cannot be right, and the answer decides which replacement styli you can buy.
- Does it need a battery, and can I change it?
- It is wireless, and the indicator panel includes a white lamp for battery low voltage, so a cell exists inside. The listing never states the cell type, the running time, or whether the user can replace it. Ask, because a sealed cell turns a consumable into a return.
- Which CNC control do the X4.3 and X4.5 addresses belong to?
- The wiring diagram gives X4.3 for probe error, X4.4 with no label, X4.5 for low battery, X4.7 for skip, and Y2.0 for the trigger, but never names the control family. On a different control you map the six cores to free inputs yourself, and you need to know the signal voltage and polarity to do it.